ATP

00:00:00 ◼   ► You think if we wait another decade, this 90s musical will wrap around and become awesome,

00:00:06 ◼   ► like the 80s music did, or will that just never happen?

00:00:12 ◼   ► As I always do, I listened to this show last week, and I kept waiting for myself to make

00:00:16 ◼   ► a point that I meant to make during the show, and it's conceivable that I stopped paying

00:00:21 ◼   ► attention for three seconds and missed myself saying this, so if I'm repeating something

00:00:26 ◼   ► I said in the very last show, I apologize.

00:00:29 ◼   ► So to be clear, you are following up with your theoretical past self.

00:00:33 ◼   ► Yeah, like I mean, I didn't have, you know, I don't have notes or anything, but when we

00:00:36 ◼   ► were talking about PSEL, I'm like, this is one of the major points I want to make, and

00:00:40 ◼   ► it's about scaling, which we also talked about in the last show, and I don't know how I could

00:00:44 ◼   ► have possibly managed to miss it.

00:00:46 ◼   ► But just in case, anyway, one of the innovations that the company that makes the PSEL thing

00:00:53 ◼   ► is supposedly bringing to the table is the ability to solve this problem that we discussed

00:00:58 ◼   ► in the last show, we're just sort of solving for what signal needs to be sent out by all

00:01:05 ◼   ► the various towers so that the interference combines to make exactly the right signals

00:01:10 ◼   ► in exactly the right place. So not just for one phone, but for two, ten, five, a thousand,

00:01:15 ◼   ► whatever. Every single one of those, all the towers that are transmitting that could possibly

00:01:20 ◼   ► be overlapping and interfering with each other, need to do so in such a way that every individual

00:01:24 ◼   ► phone gets exactly the signal it's meant for it in exactly the spot that the phone is.

00:01:30 ◼   ► And that, as you can imagine, requires fast communication between all the nodes, but also

00:01:35 ◼   ► a lot of processing power.

00:01:36 ◼   ► And the supposed innovation that the companies bring—I think Artemis is the name of it—that

00:01:41 ◼   ► the companies bring is the ability to solve that problem.

00:01:44 ◼   ► "Hey, everybody, what signal do all these towers need to put out so exactly the right

00:01:47 ◼   ► signals go in exactly the right places?"

00:01:49 ◼   ► To solve it in real time, scaling linearly with the number of phones.

00:01:54 ◼   ► So instead of, "Oh, we can solve for two phones or three phones, but as we go up to 10 phones,

00:01:58 ◼   ► then it becomes like 100 times harder to solve.

00:02:01 ◼   ► And if we go up to 1,000 phones, it's like a million times harder to solve, and there's

00:02:03 ◼   ► no way we could do it."

00:02:05 ◼   ► Supposedly they have a system where they can just add computers in a more or less linear

00:02:09 ◼   ► fashion to support more phones within the cell area.

00:02:14 ◼   ► And so that's what allows this system to be possible, that they're saying, "We have the

00:02:20 ◼   ► computing capability to do this.

00:02:21 ◼   ► We have the communication infrastructure to do it.

00:02:23 ◼   ► And we have come up with a way to solve whatever this problem is, to solve for the output that

00:02:30 ◼   ► all these towers need to put out to make constructive interference in just the right way.

00:02:36 ◼   ► And we can do it linearly.

00:02:37 ◼   ► Because if you couldn't do it linearly, you can imagine that maybe it would work for like

00:02:41 ◼   ► a trade show or something, but for an entire city with thousands upon thousands of cell

00:02:45 ◼   ► phones, if it didn't scale linearly, you've got big problems in terms of computing power.

00:02:50 ◼   ► And so that is one of, perhaps the most important claim, that and the ability to do with existing

00:02:54 ◼   ► cell phones over existing LTE networks without requiring special hardware on the handsets.

00:02:58 ◼   ► Those two things are the two things that make P cell possible, again, according to the claims.

00:03:03 ◼   ► We'll see how it shakes out.

00:03:05 ◼   ► All right.

00:03:07 ◼   ► And the next bit of follow-up was actually for me for once.

00:03:10 ◼   ► It doesn't happen often, but it does occasionally happen.

00:03:13 ◼   ► Last episode at the end, I was discussing, or John and I, and occasionally Casey making

00:03:17 ◼   ► vinyl references, we're discussing scaling and the difference between scaling and performance.

00:03:23 ◼   ► And in it, I discussed how when I'm writing web apps, I don't use database joins. In fact,

00:03:30 ◼   ► I don't use them in local apps either, but that's more because of my framework. Anyway,

00:03:35 ◼   ► so I don't use database joins and the reason why I don't use database joins is for various

00:03:40 ◼   ► scaling concerns and basically to keep my options open for splitting up that database

00:03:44 ◼   ► in the future and also to shift as much work as possible onto the

00:03:48 ◼   ► easy and cheap to scale web/application servers

00:03:52 ◼   ► and leave the hard and expensive to scale database server with as light of a workload

00:03:56 ◼   ► as possible. Anyway, so I got a bunch of feedback on

00:04:00 ◼   ► that from a lot of people. A handful of people who agreed with me and a whole bunch of

00:04:04 ◼   ► people who disagreed. The people who agreed are typically people who had

00:04:08 ◼   ► done it before, who had worked on large

00:04:12 ◼   ► large scale applications before who had seen these exact problems and solved them in the same way.

00:04:17 ◼   ► So of course they would agree because they did it the same way themselves.

00:04:22 ◼   ► The biggest argument against my technique of not using joins was that

00:04:26 ◼   ► we've never needed that in our organization or in our project, or I've personally never needed that in my project.

00:04:33 ◼   ► So you might as well use database as much as you can and when you need it then you cross that bridge.

00:04:42 ◼   ► And to me, I wrote a quick thing on my blog,

00:04:44 ◼   ► so I'm not gonna go too far into it,

00:04:45 ◼   ► but basically I think that's kind of the wrong angle

00:04:48 ◼   ► to approach scaling questions from.

00:04:50 ◼   ► Scaling is constantly asking yourself, then what?

00:04:55 ◼   ► So, okay, you have this, so far you have all these things

00:04:59 ◼   ► on one server because that one server is enough.

00:05:01 ◼   ► But then what happens when it's not enough, then what?

00:05:04 ◼   ► And scaling well and building scalable systems

00:05:08 ◼   ► means trying to anticipate some of those

00:05:11 ◼   ► than what scenarios.

00:05:12 ◼   ► And having a decent answer for it

00:05:14 ◼   ► that's not gonna involve rewrite half your code

00:05:17 ◼   ► and comb through your whole code base

00:05:19 ◼   ► and make sure you aren't using joins anywhere anymore

00:05:22 ◼   ► or rewrite, 'cause if you think about,

00:05:25 ◼   ► not only is that a ton of work,

00:05:27 ◼   ► and it could potentially introduce tons of bugs

00:05:30 ◼   ► if you haven't written this way from the beginning,

00:05:33 ◼   ► but also you'd probably be doing that

00:05:36 ◼   ► under incredible pressure and time constraints

00:05:39 ◼   ► because you're probably doing that because you hit

00:05:41 ◼   ► an oh crap moment in your scaling and you realize,

00:05:44 ◼   ► oh wait a minute, I've added the most RAM to this

00:05:47 ◼   ► than I could possibly add and it's not helping.

00:05:50 ◼   ► Or this, I've done everything I can with one box

00:05:53 ◼   ► and it still isn't enough, what do I do?

00:05:56 ◼   ► And if you don't plan for those,

00:06:00 ◼   ► and obviously this is all qualified with,

00:06:04 ◼   ► to an extent or within reason.

00:06:06 ◼   ► You don't have to plan to be the next Facebook

00:06:08 ◼   ► because you probably won't be.

00:06:09 ◼   ► And if you do become the next Facebook,

00:06:11 ◼   ► you can hire a bunch of smarter people

00:06:12 ◼   ► that need to do all this as you get that big.

00:06:14 ◼   ► But even just going from tiny to small,

00:06:19 ◼   ► or from small to midsize,

00:06:21 ◼   ► you're gonna hit some of these questions

00:06:23 ◼   ► and you're not gonna have a giant IT staff,

00:06:26 ◼   ► you're not gonna have infinite money,

00:06:27 ◼   ► you're not gonna have infinite time.

00:06:29 ◼   ► And so it's worth considering that because,

00:06:31 ◼   ► and a lot of people would say,

00:06:34 ◼   ► oh, this is about premature optimization.

00:06:36 ◼   ► And premature optimization is wrong, period.

00:06:38 ◼   ► 'cause you've heard that before and it sounds good.

00:06:41 ◼   ► And the fact is that's,

00:06:42 ◼   ► premature optimization is not always wrong or bad.

00:06:46 ◼   ► It's a trade-off.

00:06:47 ◼   ► You're trading, in most cases,

00:06:48 ◼   ► you're trading some complexity

00:06:51 ◼   ► for some or a lot of performance.

00:06:53 ◼   ► Or in the case of scaling,

00:06:57 ◼   ► options in the future for better scaling

00:07:00 ◼   ► or easier scaling or possible scaling.

00:07:02 ◼   ► And so premature optimization

00:07:05 ◼   ► and designing things to be scalable can be worthwhile

00:07:09 ◼   ► if it's relatively easy and if there's not a lot

00:07:12 ◼   ► of major costs to it, both in time and in maintenance

00:07:16 ◼   ► and limitations in the app and other things.

00:07:19 ◼   ► Like if it's relatively easy to make better decisions

00:07:23 ◼   ► that make your app more scalable the whole time

00:07:25 ◼   ► you're designing it, you should do it

00:07:27 ◼   ► because it's like a best practice.

00:07:30 ◼   ► It isn't that much more work and you might never need it,

00:07:34 ◼   ► But when you do need it, you'll be very glad you have it.

00:07:37 ◼   ► So I guess that's it on that topic for now.

00:07:40 ◼   ► Do you guys have any other follow up on scaling?

00:07:42 ◼   ► - I think there were two other threads of feedback

00:07:45 ◼   ► on the thing.

00:07:46 ◼   ► One were the people who were saying that it sounded

00:07:50 ◼   ► like you were using MySQL without using

00:07:52 ◼   ► any relational features.

00:07:53 ◼   ► So here are a bunch of non-relational data stores

00:07:56 ◼   ► that you could use instead.

00:07:57 ◼   ► So there was a couple people who had that.

00:07:59 ◼   ► And then the other one was, I think maybe it was

00:08:01 ◼   ► just one person, but I thought it was a good point

00:08:03 ◼   ► that we didn't get to, is that depending

00:08:05 ◼   ► on the nature of your application,

00:08:07 ◼   ► if you have something that's sort of non-interactive

00:08:09 ◼   ► and trivially silo-able, you could get away with--

00:08:13 ◼   ► you wouldn't have to worry about this join type thing,

00:08:15 ◼   ► because all you would do is just sort of shard up by customer.

00:08:18 ◼   ► And one customer's data would never mingle with another's.

00:08:21 ◼   ► And at that point, then you have your entire scale,

00:08:25 ◼   ► your entire horizontal scaling strategy

00:08:27 ◼   ► is entirely based on divvying users up

00:08:28 ◼   ► into smaller and smaller bins.

00:08:30 ◼   ► And you just need some way to figure out

00:08:31 ◼   ► user's bin is, and that could be something super fast and easily scalable.

00:08:37 ◼   ► And then just the bins, it's like, well, if the bin is too slow, my solution is to cut

00:08:40 ◼   ► the bin in half.

00:08:41 ◼   ► And that point, if you do joins against the bins, it doesn't matter.

00:08:44 ◼   ► All it does is make you have to cut them into pieces sooner.

00:08:47 ◼   ► Obviously, social applications are not like this, where there's some sort of global awareness,

00:08:53 ◼   ► or you're following that person's things and they're following yours or whatever.

00:08:57 ◼   ► Most systems are not this simple, but it was a good point that it really depends on the

00:09:00 ◼   ► shape of your system. What does the state look like in your thing? What state are you

00:09:05 ◼   ► tracking? How is the state related to the other state? And if you're sure that there

00:09:11 ◼   ► is no interrelations, nor will there be in the future, you can get away with just sharding

00:09:16 ◼   ► by user.

00:09:17 ◼   ► That's a very big if, though. And I agree. If you have something like, for instance,

00:09:23 ◼   ► if you look at something like Zendesk or FreshBooks, these hosted applications that they create

00:09:30 ◼   ► a little site or thing for you as the customer. And these are, by definition, very private

00:09:35 ◼   ► things that none of their other customers ever need to access your data. In fact, it

00:09:41 ◼   ► should not be possible for them to access your data. So it makes sense when you have

00:09:45 ◼   ► something that is that strictly divided by user, where the users will never have to interact

00:09:51 ◼   ► with each other or their data. It makes sense to then do that kind of sharding there.

00:09:56 ◼   ► But if you have any kind of consumer facing

00:10:00 ◼   ► general purpose service kind of thing,

00:10:02 ◼   ► like there are so many options for cross data referencing.

00:10:07 ◼   ► And if you start out with a system

00:10:12 ◼   ► that's charted in that way,

00:10:13 ◼   ► then first of all, that is a pretty big limitation

00:10:18 ◼   ► you put in yourself on your development

00:10:19 ◼   ► and it does add complexity

00:10:20 ◼   ► and it does add management complexity.

00:10:22 ◼   ► So that's the kind of thing that I would be very careful

00:10:26 ◼   ► decide, if you're deciding that up front to do that from the beginning, I would be very

00:10:29 ◼   ► careful with that decision, because that really will limit you later, and it might not be

00:10:35 ◼   ► necessary or it might not be right, or you might have to pivot into something that has

00:10:40 ◼   ► a little bit more data crossover.

00:10:44 ◼   ► Even consumer-facing systems, though, I think real systems that are out there, a lot of

00:10:48 ◼   ► the times they will take the part of their system that scales in this way and scale it

00:10:52 ◼   ► in that way, and then use a different strategy for the interrelated, because that's what

00:10:55 ◼   ► have to end up doing when you scale things is you're going to end up breaking apart the

00:10:59 ◼   ► functionality of your application, which is part of the reason why you're not doing joins,

00:11:02 ◼   ► but just like in the grand scheme of things, like maybe user information could be sharded

00:11:06 ◼   ► up nicely by user, but then all the interrelated information has an entirely different backend

00:11:10 ◼   ► with an entirely different scaling strategy.

00:11:12 ◼   ► You end up, you know, it's the whole service-oriented architecture thing, you end up with different

00:11:16 ◼   ► pieces of your application scaling in different ways, because very often it's not, there is

00:11:19 ◼   ► no one master scaling solution for your entire app, you have to look at it in pieces, and

00:11:23 ◼   ► like the login flow is like this, session management

00:11:25 ◼   ► is like this, user information is like this,

00:11:27 ◼   ► relationship management is like this.

00:11:29 ◼   ► I mean, one of the crazy examples,

00:11:30 ◼   ► I remember I wrote a story a while back

00:11:32 ◼   ► about how LinkedIn had a fairly standard looking back end

00:11:35 ◼   ► except for the part that handled all the relationships

00:11:37 ◼   ► and that had to be on a server where it was all in memory.

00:11:40 ◼   ► And I remember reading that and feeling bad for them

00:11:41 ◼   ► 'cause it's like, you know,

00:11:43 ◼   ► that's like a built-in time bomb for scaling.

00:11:45 ◼   ► Like they're racing more as law to see

00:11:48 ◼   ► will the relationship information in LinkedIn

00:11:52 ◼   ► get big faster than we can buy machines,

00:11:54 ◼   ► faster than the amount of RAM

00:11:56 ◼   ► we can stuff into a single machine.

00:11:58 ◼   ► And so that's kind of an unfortunate thing,

00:11:59 ◼   ► but it's obvious they didn't use that same scaling strategy

00:12:01 ◼   ► for the user information,

00:12:02 ◼   ► it was just for the relationship information,

00:12:04 ◼   ► which is another system.

00:12:05 ◼   ► And then one more thing, that's the then what question.

00:12:08 ◼   ► One of the then what's that you haven't asked

00:12:11 ◼   ► and didn't have to ask for a Tumblr,

00:12:13 ◼   ► I guess probably will never have to ask

00:12:14 ◼   ► for any of your applications,

00:12:15 ◼   ► but I think a lot of people thinking about

00:12:18 ◼   ► the then what's should consider

00:12:19 ◼   ► is the idea of making database accesses from essentially a web

00:12:26 ◼   ► application, or any application for that matter,

00:12:28 ◼   ► like the code that is running your thing,

00:12:30 ◼   ► it connects to a database.

00:12:32 ◼   ► Very often in these large applications,

00:12:34 ◼   ► they need to find somewhere to scale.

00:12:36 ◼   ► They have to put a layer in between there.

00:12:38 ◼   ► So you want some kind of data access layer

00:12:40 ◼   ► that does not connect to the database to get information.

00:12:43 ◼   ► And so if you're using something like, what is it,

00:12:45 ◼   ► FC model or whatever you're using,

00:12:46 ◼   ► or any of these things that basically you're

00:12:49 ◼   ► you're putting SQL in your code somewhere, right?

00:12:52 ◼   ► And you're like, well, it doesn't matter.

00:12:53 ◼   ► You don't see the SQL.

00:12:54 ◼   ► It's all abstracted away, but essentially what you're making

00:12:56 ◼   ► is a front end to do SQL queries for you.

00:12:58 ◼   ► Like if I needed to, I could swap that out

00:13:00 ◼   ► for something that didn't talk directly to a database,

00:13:03 ◼   ► but it would be kind of painful.

00:13:05 ◼   ► And so if you make a data access layer

00:13:07 ◼   ► that is agnostic to the destination,

00:13:09 ◼   ► either you make it work over HTTP from the very start

00:13:11 ◼   ► and then have sort of a web service back end

00:13:14 ◼   ► that gets you that information, you're like,

00:13:15 ◼   ► well, that seems like it's gonna be slower

00:13:17 ◼   ► and it seems like a waste of time

00:13:19 ◼   ► and I'm never going to need to do that, maybe you're right.

00:13:20 ◼   ► But in my experience, a lot of applications,

00:13:23 ◼   ► one of the first things they run into,

00:13:24 ◼   ► even if they have perfect horizontal scalability

00:13:26 ◼   ► in terms of sharding users,

00:13:27 ◼   ► 'cause it's that type of thing.

00:13:28 ◼   ► Say it's like a B2B business where you're always,

00:13:31 ◼   ► you know, your customers do want their things to be solid.

00:13:34 ◼   ► Like, oh, it's no problem.

00:13:35 ◼   ► Well, eventually you'll run into the realization

00:13:37 ◼   ► that most, especially if you're using a relational database,

00:13:39 ◼   ► most relational database products are not made

00:13:42 ◼   ► to support the number of connections

00:13:43 ◼   ► that are anything close to sort of the scale

00:13:45 ◼   ► of users on a website.

00:13:47 ◼   ► And so you have to conserve those database connections.

00:13:49 ◼   ► You need some sort of, I mean,

00:13:50 ◼   ► they have database connection pooling,

00:13:51 ◼   ► all the other strategies, but in general,

00:13:53 ◼   ► if you divorce your application as much as possible

00:13:55 ◼   ► from where it's getting its information,

00:13:57 ◼   ► even if that information is in a database now,

00:14:00 ◼   ► having some sort of transport layer in there,

00:14:02 ◼   ► like even something as dumb as HTTP for some kind of,

00:14:04 ◼   ► again, service-oriented architecture type of thing,

00:14:06 ◼   ► that seems terrible for performance,

00:14:09 ◼   ► and often is much worse for performance.

00:14:11 ◼   ► And that's why people say, no,

00:14:13 ◼   ► Casey's probably, he was complaining about

00:14:15 ◼   ► having to make a round trip to the database for multiple--

00:14:18 ◼   ► imagine if you had to make an HTTP request or something,

00:14:20 ◼   ► and then that something would potentially go to a database.

00:14:22 ◼   ► That's even more overhead.

00:14:23 ◼   ► And yes, you are sacrificing performance and complexity

00:14:26 ◼   ► for along an axis that you think you might need to--

00:14:30 ◼   ► if that then what says, OK, well, we're

00:14:35 ◼   ► out of database connections, then what?

00:14:38 ◼   ► Or having the application connect directly to the database

00:14:42 ◼   ► is a security concern or a networking issue or whatever,

00:14:46 ◼   ► then what?

00:14:47 ◼   ► If you do everything as a service,

00:14:48 ◼   ► then you could have multiple data centers and different--

00:14:52 ◼   ► do everything over HTTP and do geographic load balancing

00:14:55 ◼   ► and have all this-- it gives you more flexibility.

00:14:57 ◼   ► So that's a then what.

00:14:59 ◼   ► That's just one more then what that I've

00:15:01 ◼   ► come across many times in my working career.

00:15:03 ◼   ► And every time I've either designed or been working

00:15:06 ◼   ► on an application that connects directly to a database,

00:15:08 ◼   ► I've regretted it.

00:15:09 ◼   ► But your mileage may vary.

00:15:11 ◼   ► - No, I mean, I agree, and I'm pretty sure,

00:15:13 ◼   ► I don't actually know this for sure,

00:15:14 ◼   ► because when I left, I actually did leave,

00:15:17 ◼   ► but I'm pretty sure at Tumblr,

00:15:20 ◼   ► that was one of the big things they did

00:15:23 ◼   ► pretty soon after I left.

00:15:24 ◼   ► Once they got a more experienced staff in there

00:15:28 ◼   ► who had worked on systems of that size,

00:15:30 ◼   ► and it was, I mean, one of the reasons I left

00:15:32 ◼   ► was that it was getting pretty far beyond

00:15:34 ◼   ► my ability to scale it myself.

00:15:35 ◼   ► And that's one of the first things they did,

00:15:39 ◼   ► was moved to that kind of architecture.

00:15:41 ◼   ► And that does make sense.

00:15:42 ◼   ► That is one of those things again, where I'd like,

00:15:45 ◼   ► I think you can go a pretty long way without doing that.

00:15:49 ◼   ► And so again, and it's a premature optimization

00:15:53 ◼   ► that you, you know, it might be worth it for you to do that.

00:15:57 ◼   ► In my case, it almost never is.

00:16:00 ◼   ► And it never, well, it never has been so far,

00:16:02 ◼   ► but certainly that is something that might be worth doing.

00:16:05 ◼   ► - I don't like that phrase though,

00:16:06 ◼   ► but it's like it built into a premature optimization,

00:16:08 ◼   ► premature is saying you're doing it sooner than you need to.

00:16:11 ◼   ► These are optimizations.

00:16:12 ◼   ► Because that's one of the vocabulary tools

00:16:16 ◼   ► wielded by programmers and arguments.

00:16:17 ◼   ► Like, well, that's a premature optimization.

00:16:19 ◼   ► It's a tautology.

00:16:20 ◼   ► You are asserting that it is premature

00:16:22 ◼   ► by calling it premature.

00:16:24 ◼   ► Why is it premature?

00:16:25 ◼   ► Is it too early or is it not too early?

00:16:27 ◼   ► That's the whole argument that we're supposed to be having.

00:16:28 ◼   ► And you can't win that argument by pulling out

00:16:30 ◼   ► the term premature optimization.

00:16:32 ◼   ► That just explains what your position is,

00:16:33 ◼   ► but then you still have to defend it.

00:16:34 ◼   ► So all these optimizations, what you're saying, Marco,

00:16:36 ◼   ► is like, they're not premature.

00:16:38 ◼   ► this is in fact exactly the right time to do it,

00:16:40 ◼   ► because my expectation is x, y, and z.

00:16:42 ◼   ► And then you're arguing about, are we ever going to get big?

00:16:45 ◼   ► As we scale, which thing will break first?

00:16:48 ◼   ► You have to basically decide, what

00:16:49 ◼   ► is our maximum possible size?

00:16:51 ◼   ► What's the first thing that's going to fall over?

00:16:53 ◼   ► How long can we go doing joins?

00:16:55 ◼   ► How long can we go directly connecting to the database

00:16:57 ◼   ► and just map them out and have arguments

00:16:59 ◼   ► about which one you think is going to come first?

00:17:01 ◼   ► And then you live and learn, and applications change shape

00:17:04 ◼   ► as you go, and inevitably you make some bad choices.

00:17:06 ◼   ► And maybe you go, oh, I thought we

00:17:07 ◼   ► were going to grow in this way.

00:17:08 ◼   ► but it turns out this feature ended up being much more popular, and we hit this thing before

00:17:11 ◼   ► we hit that thing. So you're always trying to second guess what the future is. But the

00:17:14 ◼   ► whole point is, don't do any premature optimizations. Do ones that you know are not premature, in

00:17:20 ◼   ► fact, they're exactly the right time to do them. And to know what that is, you kind of

00:17:25 ◼   ► have to guess, but you use your experience and your knowledge of building similar applications

00:17:29 ◼   ► in the past. Right, like maybe instead of thinking of them

00:17:33 ◼   ► as premature, you think of them as whether they're worthwhile or not. Anyway, and the

00:17:38 ◼   ► The only other thing I would say about having that multi-tiered architecture is, for whatever

00:17:44 ◼   ► it's worth, if database connections are your problem, you're in a weird situation I've

00:17:50 ◼   ► never seen before.

00:17:52 ◼   ► I have never seen a database be limited by its number of connections.

00:17:56 ◼   ► Maybe that's because I write my code to disconnect when it's done.

00:18:00 ◼   ► Connecting and disconnecting is another thing.

00:18:02 ◼   ► A lot of, especially old school relations databases, are not designed to handle connection

00:18:06 ◼   ► storms.

00:18:08 ◼   ► How many users can try to connect to your database once?

00:18:10 ◼   ► Say 100,000 users want to connect at the same time.

00:18:13 ◼   ► What does that do?

00:18:14 ◼   ► Versus those 100,000 users being spread out over a minute, over 30 seconds, over one second.

00:18:18 ◼   ► How does it handle that?

00:18:20 ◼   ► And then you're like, "Well, we'll have persistent connections."

00:18:22 ◼   ► Okay, but now you have people tying up connections, and now you have the number.

00:18:25 ◼   ► I see it all the time.

00:18:26 ◼   ► Yeah, fair enough.

00:18:28 ◼   ► That's probably more usage-specific with whatever your application actually is, and however

00:18:34 ◼   ► your layer actually treats the database connection.

00:18:37 ◼   ► And how good your pooling solution is, because again, relational databases, depending on

00:18:40 ◼   ► their vintage, may or may not have robust pooling solutions.

00:18:43 ◼   ► At Google scale, though, one of the fun things is you run out of port numbers.

00:18:47 ◼   ► Oh, yeah.

00:18:48 ◼   ► Right, so when they do the web sockets and they're like, you know, what do you get?

00:18:51 ◼   ► It's like a 16-bit number.

00:18:52 ◼   ► You get 65,000 of them or whatever.

00:18:54 ◼   ► And at Google scale, you run into limitations in the operating system.

00:18:59 ◼   ► You're like, well, this machine could support more connections, but we were literally out

00:19:02 ◼   ► of port numbers.

00:19:03 ◼   ► That's insane.

00:19:04 ◼   ► It makes sense, but that is not something I ever would think to be a problem.

00:19:07 ◼   ► - Yeah, well that's what Google's for,

00:19:08 ◼   ► to find those limits and patch Linux to overcome them

00:19:12 ◼   ► and figure something else out.

00:19:14 ◼   ► - Yeah, I'll say the other thing,

00:19:15 ◼   ► I never got into the level where I had to start

00:19:17 ◼   ► tweaking kernel settings and things like that.

00:19:19 ◼   ► Like we would use the kernel's stock

00:19:21 ◼   ► because that was like,

00:19:23 ◼   ► and again, I don't know what Tumblr's situation is now.

00:19:26 ◼   ► They're way bigger now,

00:19:27 ◼   ► especially now that Yahoo bought them.

00:19:28 ◼   ► They're probably getting a lot of help from Yahoo's people

00:19:31 ◼   ► and running on some of Yahoo's infrastructure, who knows.

00:19:33 ◼   ► But certainly at the time I left,

00:19:35 ◼   ► that we never had to get into that.

00:19:37 ◼   ► And yeah, anyway, we are sponsored this week

00:19:41 ◼   ► by our friends once again at HelpSpot.

00:19:43 ◼   ► Now, if you're still using email clients

00:19:45 ◼   ► for customer support,

00:19:46 ◼   ► you're probably losing track of important tickets,

00:19:49 ◼   ► trying to use Mark as Unread as an organizational tool,

00:19:51 ◼   ► and IMing coworkers to see who's working on what.

00:19:54 ◼   ► It's time to get organized.

00:19:56 ◼   ► Now, most help desk apps try to be all things to all people.

00:20:00 ◼   ► HelpSpot on the other hand is focused.

00:20:02 ◼   ► It deals only with customer inquiries

00:20:04 ◼   ► and self-serve knowledge bases.

00:20:06 ◼   ► There's no asset management, password resetting,

00:20:08 ◼   ► or other unnecessary features to get in your way

00:20:10 ◼   ► or require complex integration work.

00:20:13 ◼   ► Now help desk software is usually really expensive.

00:20:16 ◼   ► A lot of times it's around $600 per user per year,

00:20:19 ◼   ► for instance.

00:20:20 ◼   ► Now help spot is just $299 per user, one time,

00:20:24 ◼   ► and that's it, you own it for life.

00:20:26 ◼   ► With help spot, there's also no lock-in.

00:20:28 ◼   ► You can download it and host it yourself,

00:20:30 ◼   ► or you can have it hosted for you,

00:20:31 ◼   ► so you have those options available,

00:20:33 ◼   ► and either way, even if they host it for you,

00:20:35 ◼   ► you always have direct access to the database

00:20:38 ◼   ► and you can directly query it

00:20:39 ◼   ► or you can export your data and take it elsewhere.

00:20:42 ◼   ► HelpSpot has been around for a long time.

00:20:44 ◼   ► It's on some new startup.

00:20:45 ◼   ► It's been available for nearly a decade

00:20:47 ◼   ► and has been adopted by thousands of companies

00:20:49 ◼   ► and organizations.

00:20:50 ◼   ► Customers from single person startups

00:20:52 ◼   ► to Fortune 500 companies use HelpSpot

00:20:54 ◼   ► to manage their support teams.

00:20:55 ◼   ► You can start a free trial at helpspot.com/atp.

00:20:59 ◼   ► Start a free trial today

00:21:01 ◼   ► and you can get $100 off your purchase

00:21:02 ◼   ► if you use the coupon code ATP14.

00:21:05 ◼   ► Once again, free trial at helpspot.com/ATP

00:21:09 ◼   ► and get $100 off your purchase

00:21:11 ◼   ► when you use the coupon code ATP14.

00:21:13 ◼   ► Thank you very much to HelpSpot

00:21:15 ◼   ► for sponsoring our show once again.

00:21:16 ◼   ► - All right, I have a question for the both of you guys.

00:21:20 ◼   ► And I think it applies to both of you,

00:21:23 ◼   ► but I'm gonna start with you, Marco.

00:21:24 ◼   ► So how do you get outside of your comfort zone

00:21:28 ◼   ► and learn something new?

00:21:29 ◼   ► And if you think about your history,

00:21:33 ◼   ► You wrote Tumblr in PHP, you wrote Instapaper in PHP,

00:21:38 ◼   ► you used MySQL for both the magazine PHP and MySQL,

00:21:42 ◼   ► now Overcast PHP and MySQL, even Second Crack,

00:21:46 ◼   ► which was kind of a for fun pet project,

00:21:49 ◼   ► was PHP and MySQL, or well, I guess not MySQL, but PHP.

00:21:52 ◼   ► How do you learn something new?

00:21:56 ◼   ► How do you go about that?

00:21:57 ◼   ► Do you not care?

00:21:58 ◼   ► Does it not bother you?

00:21:59 ◼   ► Do you just don't think it's an issue right now

00:22:01 ◼   ► and you'll learn it when you need to?

00:22:03 ◼   ► because the only thing, the only example that I can think of knowing your history is when

00:22:06 ◼   ► you learned iOS development for Instapaper, but have, what have you done since then? And

00:22:11 ◼   ► what cued me off on this was everyone saying you should look at, you know, NoSQL databases,

00:22:17 ◼   ► and you basically said, "Well, I've got MySQL and it works fine for me."

00:22:21 ◼   ► Well, it's a combination of factors. I mean, one is certainly that I'm not very good

00:22:26 ◼   ► at making myself learn new things if I don't have to. You know, that's, that's just

00:22:30 ◼   ► a character flaw I guess. No question, that's one of the factors.

00:22:34 ◼   ► But also, there's two others. One is

00:22:38 ◼   ► conservatism. That, you know,

00:22:42 ◼   ► I've been on the cutting edge before. I have lived in the cutting edge

00:22:46 ◼   ► world before. And it's a lot of work. And it's a lot of

00:22:50 ◼   ► maintenance and versions of weird things running on your server that crash

00:22:54 ◼   ► in the middle of the night and need attention, or hitting weird bugs that

00:22:58 ◼   ► because you're running such a new version of things

00:23:00 ◼   ► or such a new technology that hasn't really

00:23:02 ◼   ► had all the bugs ironed out yet.

00:23:05 ◼   ► So that's another big part of it that I just,

00:23:08 ◼   ► I'm now at a point where I want to avoid

00:23:10 ◼   ► those kind of costs if I can.

00:23:12 ◼   ► And that's not always going to be the case.

00:23:14 ◼   ► And sometimes it's worth the cost.

00:23:16 ◼   ► So for instance, on overcast servers,

00:23:18 ◼   ► I've decided to run HHVM as the interpreter layer

00:23:23 ◼   ► because it is just so much faster than PHP.

00:23:26 ◼   ► And what I'm probably gonna do,

00:23:27 ◼   ► I was setting up my load balancing today,

00:23:30 ◼   ► and I'm probably going to have two web front-end VMs to start.

00:23:34 ◼   ► And one of them I'll run HHVM, and one of them I'll run PHP.

00:23:36 ◼   ► So if it crashes weirdly, I'll still

00:23:38 ◼   ► have something to take up the slack until I fix it.

00:23:42 ◼   ► But for whatever it's worth, generally speaking,

00:23:46 ◼   ► I try to avoid the cost of being on the bleeding edge,

00:23:50 ◼   ► because I've decided it's not worth it.

00:23:52 ◼   ► You might have a different calculation on that

00:23:54 ◼   ► for whatever your application might be.

00:23:55 ◼   ► For me, I've just decided it's not

00:23:57 ◼   ► being on the bleeding edge. Now that being said, obviously there's

00:24:01 ◼   ► a lot of things that I could learn besides PHP and MySQL and Objective-C

00:24:05 ◼   ► for iOS development. There's a lot of things I could learn that are mature

00:24:09 ◼   ► and are far from being on the bleeding edge these days. And for most of those

00:24:13 ◼   ► I haven't simply because it hasn't yet been really necessary.

00:24:17 ◼   ► Learning things has a big cost and I have to weigh whether

00:24:21 ◼   ► it's worth going through the learning period, going through the period of finding all of the

00:24:25 ◼   ► intricate details of the behavior and finding out the best ways to do things

00:24:29 ◼   ► and learning the APIs in enough depth to use them really

00:24:33 ◼   ► well. And I'm at that point with the technologies I know now for the most part.

00:24:37 ◼   ► And for me to learn a whole new language or platform

00:24:41 ◼   ► or something, it has to be worth that learning curve.

00:24:45 ◼   ► And I've decided that for what I'm doing currently it's not worth

00:24:49 ◼   ► that. And part of that, obviously like on the iOS side

00:24:53 ◼   ► I don't think there's anything I can learn that would be better than Objective-C.

00:24:56 ◼   ► I don't think there are any better things really than, you know, for what I'm doing,

00:25:00 ◼   ► I think I'm using the best things already.

00:25:02 ◼   ► For the website, certainly you can do a lot better than PHP these days.

00:25:07 ◼   ► I still would argue that that might not be the case for MySQL,

00:25:11 ◼   ► but certainly you can do a lot better than PHP.

00:25:14 ◼   ► The main reason I haven't gone there on the website

00:25:17 ◼   ► is not only that I don't really need to,

00:25:22 ◼   ► to, like nothing's forcing me, and not only do I already know it and it's

00:25:26 ◼   ► very stable, but also I just don't care that much about the website.

00:25:30 ◼   ► I am not that interested in running cutting edge web stuff or even

00:25:34 ◼   ► learning new web stuff that I don't really have to learn because the website to me

00:25:38 ◼   ► is a supporting role. I am not making a web app

00:25:42 ◼   ► that happens to have an iOS client. I'm making an iOS app that happens

00:25:46 ◼   ► to have a web backend. And that's where my focus is. I care

00:25:50 ◼   ► so much more about the iOS thing. That's why I want to be cutting edge on the iOS side.

00:25:54 ◼   ► I want to be doing everything right on the iOS side. But the

00:25:58 ◼   ► web side is really just serving a supporting role here and

00:26:02 ◼   ► it's more important for me. I don't care about new

00:26:06 ◼   ► web stuff. I'm very interested in learning new frameworks for iOS.

00:26:10 ◼   ► It's not like when so much of this cool stuff that's coming around like Grand Central Dispatch and stuff like that,

00:26:14 ◼   ► like so much of this cool stuff that's coming around in iOS, I'm really into. I love that stuff.

00:26:18 ◼   ► I find it very interesting.

00:26:19 ◼   ► - That's a fair point.

00:26:20 ◼   ► - New web stuff, I just don't care about.

00:26:22 ◼   ► I don't care about web programming.

00:26:24 ◼   ► I really don't enjoy web programming anymore.

00:26:26 ◼   ► I find it very boring.

00:26:27 ◼   ► And I do it because I have to

00:26:29 ◼   ► to build the products I want.

00:26:31 ◼   ► So the primary goal is,

00:26:34 ◼   ► I like client-side programming for iOS in C,

00:26:36 ◼   ► and Objective-C,

00:26:38 ◼   ► and I like the products that result from these things.

00:26:40 ◼   ► And so the web code base is really just a,

00:26:43 ◼   ► you know, supporting role for that.

00:26:44 ◼   ► So for me, my priority there

00:26:47 ◼   ► "Just get it done. I don't want to spend a ton of time on it. Just get it done so that it can do what I need it to do."

00:26:52 ◼   ► And that doesn't include, like, learn a bunch of new web languages when I'm

00:26:57 ◼   ► not really being presented with clear motivations to do that.

00:27:02 ◼   ► Yeah, that makes sense. And I think the

00:27:07 ◼   ► most surprising use of PHP for me isn't

00:27:12 ◼   ► or Instapaper or Tumblr, it's second crack.

00:27:15 ◼   ► And granted, it makes perfect sense

00:27:17 ◼   ► that you would use something you're familiar with,

00:27:18 ◼   ► especially for a project that's not gonna make any money

00:27:21 ◼   ► and you're just doing for the heck of it.

00:27:23 ◼   ► But when I write something outside of work

00:27:26 ◼   ► and I do it for fun,

00:27:28 ◼   ► I tend to force myself to use a different technology.

00:27:31 ◼   ► For example, I wrote FastText, which is an iOS app,

00:27:35 ◼   ► and I did that in Objective-C because I didn't know

00:27:37 ◼   ► Objective-C and I wanted to learn it.

00:27:39 ◼   ► And I have been dabbling with writing my own kind of second crack.

00:27:44 ◼   ► Let's call it third crack for the sake of a conversation.

00:27:46 ◼   ► I've been using Node because I've heard a lot of great things about Node and I wanted

00:27:49 ◼   ► to try it.

00:27:50 ◼   ► I've actually really, really come to like it.

00:27:52 ◼   ► And it's just surprising to me, perhaps because you don't have a corporate stooge

00:27:56 ◼   ► job so you have a lot more control over your own destiny, but it's surprising to me that

00:28:01 ◼   ► you don't have this thirst to learn something new.

00:28:06 ◼   ► When I don't learn a new language in about a six-month to one-year window, I start to

00:28:11 ◼   ► get kind of crabby and angsty and anxious, and I really need to learn something new.

00:28:17 ◼   ► And it seems like, for better or worse, that doesn't seem to apply to you.

00:28:21 ◼   ► Well, it does, and totally.

00:28:23 ◼   ► I get that, like, the programmer itch, basically.

00:28:26 ◼   ► I totally get that.

00:28:27 ◼   ► But I satisfy it in different areas than the web language I use.

00:28:32 ◼   ► crack, you know, for those who don't know or care, the second crack is my custom static

00:28:36 ◼   ► blogging engine that I wrote. And it's just like a bunch of command line scripts basically,

00:28:40 ◼   ► and they're all written in PHP. And I'm probably the only person in the universe using PHP

00:28:44 ◼   ► to write command line scripts, and that's fine. And you're right, that totally would

00:28:49 ◼   ► have been a great opportunity to try a new language, because it was a small project with

00:28:53 ◼   ► small needs that was not very pressing to get it done quickly and doesn't need to scale,

00:28:59 ◼   ► really, like, you're right, that would have been great,

00:29:01 ◼   ► and maybe sometime I will, you know,

00:29:04 ◼   ► rewrite that, 'cause it's pretty crappy, really,

00:29:06 ◼   ► I mean, so maybe sometime I'll rewrite that

00:29:09 ◼   ► in something else, but, you know,

00:29:11 ◼   ► for part of the reason why, as you said,

00:29:13 ◼   ► I don't have a day job, when I wrote Second Crack,

00:29:16 ◼   ► it was itself a procrastination from my actual apps.

00:29:21 ◼   ► Like, that's part of the problem when you work for yourself

00:29:23 ◼   ► is that anytime you're able to do work like that,

00:29:27 ◼   ► it's potentially taking away from your quote work,

00:29:31 ◼   ► like your official work that you're actually

00:29:32 ◼   ► getting paid for or trying to get done.

00:29:34 ◼   ► And I forget exactly which project I was procrastinating on

00:29:38 ◼   ► when I wrote Second Crack.

00:29:39 ◼   ► I'm pretty sure it was probably Instapaper,

00:29:42 ◼   ► but I'm not positive on the timing on that.

00:29:44 ◼   ► But either way, that was itself a procrastination.

00:29:47 ◼   ► And so I didn't wanna spend a lot of time on it.

00:29:49 ◼   ► So that's why I figured, let me just use PHP

00:29:52 ◼   ► because I can get it done very quickly with PHP.

00:29:55 ◼   ► I didn't want it to be a big time sink.

00:29:57 ◼   ► Yeah, that makes sense.

00:29:58 ◼   ► That being said, though, I still learn tons--

00:30:01 ◼   ► you know, as I said, I try to keep up as much as I can

00:30:03 ◼   ► with iOS APIs, especially low-level stuff.

00:30:06 ◼   ► Like, for instance, with Overcast,

00:30:10 ◼   ► I don't use AVPlayer or AVAudioPlayer.

00:30:15 ◼   ► And I'm not sure if I've said that publicly before.

00:30:17 ◼   ► I think I have, so this shouldn't be a big surprise.

00:30:19 ◼   ► But I'm playing audio files, and I'm not using AVPlayer.

00:30:24 ◼   ► And if you've ever worked on this stuff,

00:30:26 ◼   ► you should know therefore what that means I'm using raw core audio.

00:30:30 ◼   ► And that, like, this has been an amazing learning process.

00:30:34 ◼   ► And probably I didn't need to do that.

00:30:38 ◼   ► I have some reasons why I did it, but I probably didn't need to. Chances are

00:30:42 ◼   ► you could just do, you know, I could have used debut player and saved

00:30:46 ◼   ► myself a lot of time, but I wanted certain levels of control and I wanted to learn this

00:30:50 ◼   ► stuff. So I made extensive use of things like the accelerate framework

00:30:54 ◼   ► all the VDSP functions. So these functions that basically vectorize operations and use SIMD

00:31:01 ◼   ► instructions and stuff like that to really make these things awesomely fast if possible.

00:31:06 ◼   ► And I'm using that stuff all over the place in so many ways. And so I'm learning things like that.

00:31:10 ◼   ► I'm learning various different concurrency strategies and doing things like the crazy

00:31:17 ◼   ► ring buffer for the buffering of the samples that I'm reading from the files and everything.

00:31:22 ◼   ► That stuff I love.

00:31:23 ◼   ► That's all really interesting to me.

00:31:25 ◼   ► This is all, you know, that's where I've been learning

00:31:28 ◼   ► is in overcast client-side code.

00:31:31 ◼   ► And again, the web stuff I couldn't possibly care less about.

00:31:34 ◼   ► - So John, what about you?

00:31:35 ◼   ► Because your day job has been, as far as I know,

00:31:38 ◼   ► pretty much forever in Perl.

00:31:41 ◼   ► Do you ever have that itch?

00:31:43 ◼   ► What do you do to scratch it, if so?

00:31:46 ◼   ► - Well, with less time for like fun programming projects

00:31:50 ◼   ► for all the things that I've put on my own plate,

00:31:52 ◼   ► essentially, besides my job over the years,

00:31:54 ◼   ► whether it be writing for Ars Technica,

00:31:56 ◼   ► which I used to do much more of,

00:31:57 ◼   ► or podcasting more recently.

00:31:59 ◼   ► There's not as much time for fun programming projects.

00:32:01 ◼   ► I wrote my own little static blogging engine thing,

00:32:04 ◼   ► but that was kind of like Marco's,

00:32:06 ◼   ► and they were like,

00:32:07 ◼   ► yeah, it's supposed to be kind of a fun project,

00:32:08 ◼   ► but really it was a practical thing.

00:32:10 ◼   ► I just wanted to get it done quickly,

00:32:13 ◼   ► so I wrote it in Perl

00:32:14 ◼   ► 'cause I knew we could get it done quickly,

00:32:16 ◼   ► and I intentionally stopped myself from making it good,

00:32:19 ◼   ► Because it was like, oh, because I, you know,

00:32:21 ◼   ► this is something I know how to do.

00:32:23 ◼   ► I could make this thing have bells and whistles

00:32:25 ◼   ► and be awesome and have this cool feature.

00:32:26 ◼   ► And I was like, no, make it as dumb as possible.

00:32:28 ◼   ► Get done with it.

00:32:29 ◼   ► Like, because I was, at that point I was trying

00:32:31 ◼   ► to not distract myself from writing,

00:32:32 ◼   ► which I'm not distracting myself from writing anymore

00:32:35 ◼   ► by making the engine, but I'm also not writing.

00:32:36 ◼   ► But anyway, back then it was like,

00:32:37 ◼   ► don't spend time writing the engine.

00:32:39 ◼   ► You do this all day, it's stupid.

00:32:42 ◼   ► Make the dumbest thing you can possibly do that works.

00:32:44 ◼   ► And I did, and it's super dumb,

00:32:46 ◼   ► and it didn't stop me from writing.

00:32:49 ◼   ► But the thing that I think has made me like try new things and stuff or whatever

00:32:53 ◼   ► is kind of the brutal nature of the tech job market starting in the mid to late

00:32:59 ◼   ► nineties in that I've had a lot of jobs.

00:33:01 ◼   ► And, uh, you know, like, how did I get to use, uh, MSQL, MySQL, SQLite, Oracle,

00:33:09 ◼   ► Informix, Postgres, like you're not going to get experience in all those databases,

00:33:14 ◼   ► probably in one place, even if you're there from the beginning, when it's like

00:33:17 ◼   ► three guys, maybe you'll touch on one or two of them.

00:33:20 ◼   ► But how will you ever take an in-depth survey

00:33:24 ◼   ► of all these different database products?

00:33:26 ◼   ► Well, you do it by having seven different jobs.

00:33:29 ◼   ► And the first place you work, there is no database,

00:33:31 ◼   ► so you get to pick something.

00:33:32 ◼   ► The second place maybe uses MySQL,

00:33:33 ◼   ► and the third place you decide to use Postgres

00:33:34 ◼   ► 'cause you think that might be fine.

00:33:36 ◼   ► When you make commitments like that,

00:33:37 ◼   ► they tend to last for a long time.

00:33:40 ◼   ► Or if you're some big company

00:33:41 ◼   ► that are already using Oracle,

00:33:42 ◼   ► you're not gonna come in there and go,

00:33:43 ◼   ► "You know what, I'd really like to try out Postgres."

00:33:46 ◼   ► It's like, well, we're using Oracle, so get a different job.

00:33:51 ◼   ► And languages is the same way.

00:33:52 ◼   ► If you're a web developer, yes, every place I've worked has used Perl in some fashion,

00:33:56 ◼   ► but the proportion of my time spent writing Perl is very wildly.

00:34:00 ◼   ► For example, the company where I used Postgres extensively, a huge amount of the code was

00:34:04 ◼   ► in stored procedures in Postgres, and at various points, I was mostly a Postgres database designer/stored

00:34:12 ◼   ► procedure writer.

00:34:13 ◼   ► And in FormX, we also use a lot of stored procedures.

00:34:16 ◼   ► And if you're writing an application that is mostly,

00:34:18 ◼   ► essentially, a JavaScript application that just talks

00:34:21 ◼   ► to a faceless back end, the faceless back end

00:34:23 ◼   ► is pretty boring.

00:34:25 ◼   ► And the entire application lives on the client side.

00:34:28 ◼   ► And so at that point, you're writing a JavaScript

00:34:30 ◼   ► application.

00:34:31 ◼   ► It's not Node where it's server-side JavaScript,

00:34:33 ◼   ► but client-side JavaScript with all the frameworks.

00:34:34 ◼   ► Like, well, then what framework are you using?

00:34:36 ◼   ► How did I get to use all these different frameworks,

00:34:38 ◼   ► going from plain old DOM to like prototype, Dojo, jQuery,

00:34:42 ◼   ► Angular, Ember, Backbone, Underscore.

00:34:46 ◼   ► How do you get to use all those?

00:34:47 ◼   ► You're not going to get to use all of them at one place.

00:34:49 ◼   ► Because once you start building an application on top

00:34:51 ◼   ► of one of those, you're probably going

00:34:52 ◼   ► to keep building it that way.

00:34:53 ◼   ► And then you say, oh, well, jQuery is really popular now.

00:34:56 ◼   ► So we've really picked the wrong horse

00:34:58 ◼   ► when we decided to do everything with prototype.

00:35:00 ◼   ► Rails did that a little bit to some degree as well.

00:35:02 ◼   ► And it's like, well, how many thousands of lines

00:35:05 ◼   ► are in this application now?

00:35:06 ◼   ► Do you want to rewrite it with jQuery,

00:35:08 ◼   ► or do we just continue plowing bravely onward?

00:35:10 ◼   ► And so, yeah, changing jobs, usually not voluntarily,

00:35:13 ◼   ► but more like when the company starts to go down the tubes

00:35:17 ◼   ► or have layoffs or changes direction or whatever,

00:35:21 ◼   ► you get a new job, that's a new opportunity

00:35:22 ◼   ► to learn a lot of new things.

00:35:23 ◼   ► Even if the whole time, like, I'm just a web developer.

00:35:26 ◼   ► What does web developer mean?

00:35:27 ◼   ► Web developer means being multilingual,

00:35:30 ◼   ► learning, and even CSS.

00:35:31 ◼   ► Like, well, I've got no CSS, I've got no HTML,

00:35:33 ◼   ► I've got to keep all the specs,

00:35:34 ◼   ► but there's all the languages on top of CSS,

00:35:35 ◼   ► like LESS and SAS and stuff.

00:35:37 ◼   ► you will find yourself learning new languages all the time just to keep up with the constant

00:35:42 ◼   ► amazing churn in web technologies.

00:35:45 ◼   ► And you will build products on whatever the web technologies are at the time that your

00:35:50 ◼   ► company chooses or you choose or whatever, and then when you go to the next job there'll

00:35:53 ◼   ► be a different set of technologies.

00:35:54 ◼   ► Not that I enjoy moving jobs, I don't.

00:35:55 ◼   ► I wish I could stay in one job for, you know, I don't like switching jobs, I find it stressful,

00:36:00 ◼   ► but the plain fact is that I have switched jobs many times and I think that has forced

00:36:03 ◼   ► me more than anything to learn new technologies.

00:36:06 ◼   ► That makes sense.

00:36:07 ◼   ► I'm a little surprised that neither of you seems to dedicate the time outside of work

00:36:15 ◼   ► to do this, but let's also not lose sight of the fact that you have children and I do

00:36:18 ◼   ► not. So that probably relates more than just a little bit.

00:36:22 ◼   ► I mean, I read a lot about it. I've never written a significant program in Objective-C,

00:36:28 ◼   ► but I feel like I know the language. And I've never written anything with any of these frameworks,

00:36:32 ◼   ► but I know a lot about them. And it's like, yes, it's kind of like a dilettante. It's

00:36:36 ◼   ► of like, "Well, I just read a lot of books about it." It's true. I read a lot of articles,

00:36:39 ◼   ► read a lot of books, read a lot of Ruby Python code, read a lot of Node.js code, read headers

00:36:44 ◼   ► from Cocoa to see what it's like. I'm not coding in those, but I feel like by surveying

00:36:49 ◼   ► them, I don't know if it's the same thing as like, it's certainly not the same thing

00:36:53 ◼   ► as doing a real project in it, but you kind of get a sense for what's out there, and so

00:36:56 ◼   ► it's not entirely alien to it. Like, why am I reading these big long Rust tutorials? Why

00:37:00 ◼   ► did I read all the documentation about Go? Because I have a project in mind because I'm

00:37:04 ◼   ► going to use them? Probably not, but I have an academic interest in that type of thing,

00:37:07 ◼   ► and I think reading through it gives you an idea of what it might be like to program in that.

00:37:11 ◼   ► [laughs]

00:37:11 ◼   ► Fair enough.

00:37:13 ◼   ► Actually, I did look at Go recently, because everyone kept telling me I should. It does look

00:37:18 ◼   ► really interesting, and I think if I were to go through the trouble of learning a whole new

00:37:22 ◼   ► web language, that might be the one I would go to, because it seems like it's made with my

00:37:26 ◼   ► sensibilities pretty well aligned. So, anyway. This would be a great opportunity to talk about

00:37:33 ◼   ► our second sponsor, which is lynda.com. That is L-Y-N-D-A dot com. lynda.com offers thousands

00:37:40 ◼   ► of online video courses in software, creative, and business skills. Everything from iOS and

00:37:44 ◼   ► web development to things like Photoshop or Lightroom. They have over 2,400 courses taught

00:37:49 ◼   ► by industry experts. lynda.com works with software companies to provide you with updated

00:37:54 ◼   ► training the same day a new version hits the market, so you're always up to speed. And

00:37:58 ◼   ► they have courses for all experience levels, whether you're a beginner or an advanced user

00:38:02 ◼   ► of these things. Now you get all of this, 2,400 courses, one low monthly price gets

00:38:08 ◼   ► you all of this. It's a flat rate. $25 a month for unlimited access to the entire lynda.com

00:38:13 ◼   ► library. Now, I mean, I've seen some of their videos, I've actually learned a lot from them.

00:38:20 ◼   ► For my podcast editing, I went through some of their videos on Logic and audio engineering,

00:38:25 ◼   ► and it's great. I mean, you've got to see this. You know, in case you were telling us

00:38:28 ◼   ► as last time, you gotta see how they have the video

00:38:31 ◼   ► and then they have next to it, they have a transcript

00:38:34 ◼   ► and it tracks and the transcript shows you,

00:38:36 ◼   ► it highlights where you are, what's being said.

00:38:38 ◼   ► So you can, and then you can scroll through the transcript

00:38:40 ◼   ► to jump to any point in the video.

00:38:41 ◼   ► So you can skim through and you can see,

00:38:43 ◼   ► oh, I wanna hear about this topic right here

00:38:45 ◼   ► on the transcript.

00:38:46 ◼   ► Click on that, it goes to that point in the video.

00:38:48 ◼   ► Really well done.

00:38:49 ◼   ► You can watch them on your iPad and everything.

00:38:51 ◼   ► It's fantastic.

00:38:52 ◼   ► They don't need Flash or any stupid stuff like that.

00:38:55 ◼   ► Now, relative to our recent discussions,

00:38:58 ◼   ► They have courses in web development.

00:39:01 ◼   ► For instance, that you can learn Node.js,

00:39:04 ◼   ► you can learn AngularJS or CoffeeScript,

00:39:07 ◼   ► or as many people would tell me to learn,

00:39:09 ◼   ► NoSQL databases.

00:39:11 ◼   ► They have all these things.

00:39:13 ◼   ► There's a whole course they recently added

00:39:16 ◼   ► called Programming for iOS 7 for Non-Programmers.

00:39:20 ◼   ► So you can learn how to make an app.

00:39:22 ◼   ► If you're not a programmer,

00:39:23 ◼   ► there actually are about two or three of you out there

00:39:25 ◼   ► who listen to this show who can tolerate us

00:39:26 ◼   ► and yet who aren't programmers.

00:39:28 ◼   ► You three people can finally learn how to program

00:39:31 ◼   ► if you want to.

00:39:32 ◼   ► Programming for non-programmers for iOS 7,

00:39:35 ◼   ► this great course on lynda.com.

00:39:37 ◼   ► You can even, they recently,

00:39:39 ◼   ► so recently Adobe launched Lightroom for mobile,

00:39:42 ◼   ► Lightroom for iPad basically.

00:39:44 ◼   ► And they have a course on that,

00:39:45 ◼   ► so you can learn how to edit your photos

00:39:46 ◼   ► and manage them on the go.

00:39:48 ◼   ► So this is all really fantastic stuff.

00:39:51 ◼   ► It's high quality stuff made by people

00:39:53 ◼   ► at the top of their fields.

00:39:54 ◼   ► These are not like the little homemade YouTube videos you see.

00:39:59 ◼   ► Really fantastic quality, top-notch content.

00:40:01 ◼   ► I've used this myself.

00:40:03 ◼   ► We all said we'd use this ourselves.

00:40:05 ◼   ► I really can't recommend them enough.

00:40:06 ◼   ► Fantastic.

00:40:07 ◼   ► So thanks a lot to lynda.com, L-Y-N-D-A dot com.

00:40:11 ◼   ► If you go to lynda.com/ATP, you can start a free seven-day trial.

00:40:17 ◼   ► That is L-Y-N-D-A dot com slash ATP.

00:40:20 ◼   ► Thanks a lot to Lynda for sponsoring our show once again.

00:40:23 ◼   ► Apple did something a little un-Apple-like today, and they've announced an OS X beta

00:40:29 ◼   ► program for regular people, for non-developers.

00:40:35 ◼   ► This is surprising to me.

00:40:38 ◼   ► What did you guys think?

00:40:39 ◼   ► I wasn't so old I could remember definitively what the details of the old Apple Free developer

00:40:47 ◼   ► program were, because Apple's developer program has had changed over the years.

00:40:53 ◼   ► Right now it's like $99 for the iOS dev program, $99 for the OS X dev program.

00:40:59 ◼   ► And those prices were a step down from the old program, which was like, you could be

00:41:04 ◼   ► a select developer for like $500 a year, you could be a premier developer for like $1200,

00:41:09 ◼   ► and there was no iOS, or maybe there was at the tail end of that, but anyway, there weren't

00:41:14 ◼   ► separate programs for iOS and OS X.

00:41:17 ◼   ► But there was also a free tier where you could be a developer at the free tier, and the things

00:41:21 ◼   ► I can't remember at the free tier whether you got early seeds of the early OS 10 betas

00:41:27 ◼   ► or not.

00:41:28 ◼   ► Yeah, so someone can write in and tell me the exact details of what the free programs

00:41:31 ◼   ► were like.

00:41:32 ◼   ► But anyway, this is kind of a return to the free thing.

00:41:34 ◼   ► It's like, hey, you can be a registered developer, kind of, sort of, not really, at Apple.

00:41:39 ◼   ► But it's not part of the, like, if you're just interested in getting the pre-release

00:41:41 ◼   ► software, sign up for free.

00:41:43 ◼   ► You don't have to pay $99, because I'm assuming you're not going to write your own applications.

00:41:47 ◼   ► You just want to get the early versions of the OS, which a lot of people did back in

00:41:50 ◼   ► the day.

00:41:51 ◼   ► You've got to be part of the dev program

00:41:52 ◼   ► if you want early releases, even if you're not going to make

00:41:56 ◼   ► your own application.

00:41:56 ◼   ► Which is essentially why I'm always

00:41:58 ◼   ► a member of the Mac developer program.

00:42:00 ◼   ► I'm not writing Mac applications,

00:42:01 ◼   ► but I've been a member of the Mac developer program,

00:42:05 ◼   ► a paying member for just years and years now.

00:42:07 ◼   ► Because that's how you get the early releases.

00:42:09 ◼   ► And this is an example of the ability

00:42:11 ◼   ► to get the early release, and I assume nothing else except for

00:42:14 ◼   ► just the early releases, like no developer technical support

00:42:18 ◼   ► incidents, no ability to sign apps and upload them to the App

00:42:21 ◼   ► Store and all that other stuff.

00:42:23 ◼   ► But anyway, for regular people.

00:42:26 ◼   ► So you do need an Apple ID.

00:42:27 ◼   ► You do need to agree to a confidentiality thing, where

00:42:30 ◼   ► you're not going to talk about the features that you see

00:42:32 ◼   ► and so on and so forth.

00:42:33 ◼   ► You do agree to let Apple collect diagnostics

00:42:36 ◼   ► from your computer and all the other normal stuff

00:42:38 ◼   ► that you might imagine agreeing to if you've decided

00:42:40 ◼   ► to run beta software to help out Apple.

00:42:45 ◼   ► And in return, you get to see an early peak at things.

00:42:48 ◼   ► So are you surprised by this?

00:42:51 ◼   ► Not really.

00:42:52 ◼   ► Because first of all, these people are not going to--

00:42:56 ◼   ► I assume they're not going to get the same seeds

00:42:58 ◼   ► that developers get.

00:43:00 ◼   ► Probably Apple's going to take their time and release maybe--

00:43:02 ◼   ► developers maybe get five or six seeds.

00:43:04 ◼   ► And of those, Apple will pick, here's

00:43:06 ◼   ► one seed that everybody gets.

00:43:07 ◼   ► And then there'll be a bunch of dev seeds.

00:43:09 ◼   ► And then one seed that everybody gets.

00:43:10 ◼   ► Then a bunch of dev seeds.

00:43:11 ◼   ► But nobody's getting anything until someone

00:43:13 ◼   ► goes on stage at WWDC and shows the new features and stuff.

00:43:16 ◼   ► It's not as if you will be seeing some secret thing

00:43:19 ◼   ► before everyone else.

00:43:20 ◼   ► Everyone's going to see it on stage at WWDC.

00:43:22 ◼   ► There's going to be a build at WWDC.

00:43:24 ◼   ► And at that point, it's like, well, who cares?

00:43:28 ◼   ► If you get the WWDC build, you get

00:43:30 ◼   ► the build a couple weeks after that,

00:43:31 ◼   ► like the cat's out of the bag.

00:43:32 ◼   ► So don't think that by signing up for this, wow,

00:43:34 ◼   ► I don't have to pay any money, and I

00:43:35 ◼   ► get to know all of Apple's secrets.

00:43:36 ◼   ► You're not going to know any secrets.

00:43:39 ◼   ► Continue to go to the rumor sites

00:43:40 ◼   ► if you want to see fuzzy screenshots of supposed

00:43:43 ◼   ► release things.

00:43:44 ◼   ► This is merely a way to let you, you know, it's basically, Apple is doing this because

00:43:49 ◼   ► I think they want wider testability of their OS and they're trying it out on the Mac because

00:43:53 ◼   ► that probably seems like a safer bet.

00:43:55 ◼   ► They just want more people to try stuff out so they're not surprised when they release

00:43:59 ◼   ► something to the public because there's a very small number of developers and they have

00:44:04 ◼   ► a, probably a less diverse set of hardware than the market at large.

00:44:08 ◼   ► And so Apple would prefer, I think, to get wider testing of its OS.

00:44:12 ◼   ► And not right away, in just certain builds, and you know, like, we think, you know, maybe

00:44:18 ◼   ► two or three releases leading up to it, these are the ones that we want to test widely to

00:44:21 ◼   ► check for, you know, driver compatibility issues or hardware issues or whatever.

00:44:26 ◼   ► That's why Apple's doing this, because Apple thinks it will benefit to them and make their

00:44:28 ◼   ► software more robust.

00:44:30 ◼   ► It's not them doing you a favor to let you see pre-release software.

00:44:33 ◼   ► is entirely, hey, we really need a wider testing base.

00:44:38 ◼   ► At a certain point in the development of our OS,

00:44:40 ◼   ► it comes to the point where we say,

00:44:42 ◼   ► we know all we can know with our small set of internal testing

00:44:46 ◼   ► in QA and our small set of developers who are using this.

00:44:49 ◼   ► If anyone wants to sign up for a wider beta,

00:44:52 ◼   ► we would love to know if something's

00:44:54 ◼   ► crashing on some obscure configuration

00:44:56 ◼   ► that we don't have here, and then we can collect the data

00:44:58 ◼   ► and figure it out.

00:44:59 ◼   ► And maybe they'll consider doing that with iOS eventually

00:45:02 ◼   ► as well.

00:45:02 ◼   ► but I think the Mac is like a safe place

00:45:04 ◼   ► to try out this type of program.

00:45:06 ◼   ► - Yeah, I tend to agree.

00:45:08 ◼   ► I thought that there were a couple of interesting things

00:45:13 ◼   ► that came from this.

00:45:14 ◼   ► Firstly, something that occurred to me was,

00:45:17 ◼   ► it's giving, or it's Apple participating

00:45:20 ◼   ► in a little bit more outreach

00:45:22 ◼   ► and a little bit more community involvement.

00:45:24 ◼   ► I know, especially after Microsoft Build,

00:45:27 ◼   ► there's been a lot of comparison

00:45:29 ◼   ► between Microsoft's approach to developers

00:45:32 ◼   ► and Apple's approach to developers.

00:45:33 ◼   ► Now granted, this isn't a developer specific program,

00:45:37 ◼   ► and that's kind of what I'm driving at,

00:45:39 ◼   ► is that you would think if Apple was going to reach out

00:45:42 ◼   ► into the public at all, in any capacity,

00:45:45 ◼   ► that perhaps it would be to be a little more friendly

00:45:48 ◼   ► to developers.

00:45:49 ◼   ► I agree with what you said, John,

00:45:50 ◼   ► that this is really about Apple doing what's good for Apple,

00:45:53 ◼   ► as Apple always does,

00:45:55 ◼   ► but it just seems kind of unfortunate

00:45:57 ◼   ► that they're reaching out to the wrong group.

00:45:59 ◼   ► And the other thing I was thinking about was, what if they've learned from iOS 7?

00:46:07 ◼   ► And iOS 7, even though a lot of the nerds knew it was going to look really different

00:46:11 ◼   ► and really flat, and that's really how you would describe it, but you know what I mean.

00:46:16 ◼   ► But it has depth.

00:46:17 ◼   ► Right, it has depth, of course.

00:46:19 ◼   ► And clarity.

00:46:21 ◼   ► And chamfered edges.

00:46:22 ◼   ► Oh, wait.

00:46:23 ◼   ► But anyway, what if Craig Hockenberry is right?

00:46:27 ◼   ► He wrote a really great post.

00:46:28 ◼   ► I think it was a post.

00:46:29 ◼   ► it was a series of tweets, but either way, he wrote something recently saying, "Hey,

00:46:33 ◼   ► there's a lot of indication that OS X is going to look pretty different in the next version."

00:46:39 ◼   ► And his point was, "Developers, you should really start testing with some of the fonts

00:46:43 ◼   ► we expect them to use," and so on and so forth.

00:46:45 ◼   ► Well, what if this is queuing up, or setting up for, is a better way of phrasing it, a

00:46:51 ◼   ► little bit more outreach to the public at large?

00:46:54 ◼   ► So at WWDC, we see the new OS X.

00:46:57 ◼   ► It looks really, really different.

00:47:00 ◼   ► And rather than nobody getting to see it

00:47:02 ◼   ► and nobody getting to play with it except developers

00:47:05 ◼   ► or those willing to pay the $100,

00:47:07 ◼   ► maybe they're trying to reach out

00:47:09 ◼   ► to some of the super Apple nerds

00:47:11 ◼   ► or maybe even like self-appointed evangelists, if you will,

00:47:14 ◼   ► and say, "Hey, you try it out

00:47:17 ◼   ► "and you spread the word that this isn't so bad."

00:47:19 ◼   ► I'm standing on a whole bunch of theory here,

00:47:22 ◼   ► but I don't know, it makes sense to me.

00:47:24 ◼   ► Well, it's not so much that they want them to be evangelists for them.

00:47:27 ◼   ► I think it's more like, to use an example from OS X instead of the iOS 7 thing, it's

00:47:31 ◼   ► kind of like their attempts to handle autosave and getting rid of the save command.

00:47:35 ◼   ► By sending out what's probably going to be a fairly significant UI overall of the OS

00:47:40 ◼   ► to a wider group of people, they hope to find out earlier rather than later that, oh, everybody

00:47:45 ◼   ► hates the new autosave and people want save as to be back and they want it to have a command

00:47:48 ◼   ► key.

00:47:49 ◼   ► thing you could have found out in a wider early beta, and that they only found out after

00:47:56 ◼   ► the release and had to patch up in Mountain Lion. You know what I mean? Again, it's serving

00:48:01 ◼   ► Apple. It's not just, "Hey, let's see if your thing crashes," or whatever, but also if people

00:48:06 ◼   ► flip out. Because when all those people who are looking at rumors or screenshots see something

00:48:11 ◼   ► like that, they may have trepidation about it, but Apple has no way to get feedback from

00:48:15 ◼   ► them. And really, a lot of them will think, "Well, I've got to try it and see what it's

00:48:18 ◼   ► like when I use it. But if all those people who are so enthusiastic that are reading a

00:48:21 ◼   ► rumor site had the actual pre-release of the OS and got to try it and said, "Oh, I can't

00:48:26 ◼   ► tell what's what." Like, for example, these buttons without borders, I can't tell whether

00:48:30 ◼   ► they're buttons, they just look like text or, you know, that's the iOS equivalent. But

00:48:34 ◼   ► if there's some big UI change that they're not so sure about, if they put it out to a

00:48:39 ◼   ► wider audience, they will get, A, they'll get informed feedback instead of like, "I

00:48:42 ◼   ► saw a screenshot that looks like it's going to be terrible." And B, presumably through

00:48:46 ◼   ► the mechanisms that they're going to distribute, there will be official channels to send that

00:48:49 ◼   ► feedback, not just complaining on Twitter or posting in web forums or whatever, but

00:48:54 ◼   ► actual feedback directly to Apple where people can write 20-paragraph missives about why they

00:48:59 ◼   ► don't like the new look and how they can't tell which window is in the front anymore

00:49:01 ◼   ► or whatever they're going to say. You know what I mean?

00:49:04 ◼   ► Yeah, that makes sense. I don't know. Marco, any thoughts on this?

00:49:09 ◼   ► I'm less excited about it than most people. I mean, first of all, it's mostly because

00:49:12 ◼   ► because I have zero interest in running a beta version

00:49:16 ◼   ► of OS X on my Mac, zero.

00:49:18 ◼   ► Like the only time I ever installed a beta

00:49:22 ◼   ► was when reading list was first added to it.

00:49:26 ◼   ► And I was tipped off that this was a thing in the betas

00:49:28 ◼   ► and I should see it 'cause at the time

00:49:29 ◼   ► I still owned Instapaper.

00:49:31 ◼   ► And so I signed up with the developer program,

00:49:34 ◼   ► paid the 100 bucks to see this reading list thing

00:49:37 ◼   ► and see how worried I had to be basically.

00:49:42 ◼   ► Even then, I installed it on a laptop that I didn't care that much about because it's

00:49:46 ◼   ► so important to me that my computer work perfectly, as much as possible.

00:49:51 ◼   ► And whenever my computer does not work perfectly, as much as possible, it's very disruptive

00:49:57 ◼   ► and it's potentially very costly to me.

00:49:59 ◼   ► And so I always want to make sure that I'm okay in that regard.

00:50:06 ◼   ► And so it's just not worth me running the betas, for the most part, ever.

00:50:11 ◼   ► I do it on iPhone because the iPhone is not that important to me.

00:50:13 ◼   ► If the iPhone, you know, if my iPhone reboots twice a day, that's annoying, but, you know,

00:50:21 ◼   ► oh well.

00:50:22 ◼   ► And it's more important for me in iOS because my software runs on iOS.

00:50:25 ◼   ► It's more important for me to get into that as early as possible and to have my software

00:50:29 ◼   ► running on my main phone.

00:50:31 ◼   ► When something like iOS 7 comes out that changes a whole bunch about the entire paradigm of

00:50:35 ◼   ► the UI, the entire style, it's good to start getting that ingrained in me earlier so I

00:50:40 ◼   ► so I can develop for it well, things like that.

00:50:43 ◼   ► But because it doesn't apply to me on the OS X side,

00:50:48 ◼   ► I just don't care at all.

00:50:49 ◼   ► Now, maybe the problem was not enough people

00:50:53 ◼   ► were testing things on the OS X side

00:50:55 ◼   ► to make the testing really worthwhile to Apple

00:50:58 ◼   ► or to accomplish their goals.

00:51:00 ◼   ► 'Cause you know, do you see them doing this for iOS?

00:51:03 ◼   ► I sure as hell don't.

00:51:04 ◼   ► Because the iOS beta is like,

00:51:06 ◼   ► everyone who wants the iOS beta gets it.

00:51:08 ◼   ► You know, they find some developer

00:51:10 ◼   ► to get them to add it to their account,

00:51:11 ◼   ► and to add their device to their account,

00:51:13 ◼   ► and they get the beta.

00:51:14 ◼   ► Or everyone pulls together money

00:51:16 ◼   ► and buys a slot on someone's,

00:51:18 ◼   ► there's all these weird things people do

00:51:20 ◼   ► to get iOS betas these days.

00:51:22 ◼   ► And the reason they don't really need to do it there,

00:51:26 ◼   ► and they probably won't do it there,

00:51:28 ◼   ► is because there's all this demand to run iOS betas.

00:51:31 ◼   ► All that demand probably isn't really there

00:51:32 ◼   ► for the OS X betas.

00:51:34 ◼   ► And part of that's because Macs

00:51:36 ◼   ► aren't as interesting as iPhones.

00:51:38 ◼   ► There are a lot fewer people who have them.

00:51:40 ◼   ► And part of that's just because OS X

00:51:41 ◼   ► doesn't change that much.

00:51:43 ◼   ► Maybe this next beta will be so different,

00:51:46 ◼   ► you know, maybe 10.10 will be so different

00:51:49 ◼   ► that everyone will want it, who knows.

00:51:51 ◼   ► But I'm guessing they weren't getting enough people testing.

00:51:54 ◼   ► And that's why you had things like

00:51:56 ◼   ► the crazy Gmail problems in Mavericks

00:51:58 ◼   ► that like, why wasn't that caught in a beta?

00:52:00 ◼   ► You know, stuff like that.

00:52:01 ◼   ► Because I'm guessing most people who are on the OS X betas

00:52:05 ◼   ► probably have the same priorities I do

00:52:09 ◼   ► in that they probably don't wanna run them

00:52:10 ◼   ► on their main work machines.

00:52:12 ◼   ► So they probably are like developers who make Mac apps

00:52:16 ◼   ► who wanna make sure their Mac apps work.

00:52:18 ◼   ► So they probably run them in a VM or on some other hardware

00:52:21 ◼   ► that that's not their primary stuff.

00:52:23 ◼   ► And then they just test their app.

00:52:25 ◼   ► So if that's what most of the testers are doing,

00:52:27 ◼   ► it's in Apple's best interest

00:52:28 ◼   ► to try to broaden that tester base.

00:52:31 ◼   ► But again, I don't see this as that big of a deal.

00:52:33 ◼   ► I don't see how many people who are not developers

00:52:36 ◼   ► and/or who were unwilling to pay

00:52:38 ◼   ► for the developer account before,

00:52:39 ◼   ► I don't see how many people are gonna really get

00:52:42 ◼   ► who are gonna be dying to install

00:52:44 ◼   ► what's probably going to be a boring beta

00:52:47 ◼   ► of a boring OS on boring products

00:52:49 ◼   ► because they're all pretty mature at this point.

00:52:50 ◼   ► - Yeah, speaking of the Gmail thing,

00:52:52 ◼   ► that's one aspect of,

00:52:55 ◼   ► even if you do install the betas,

00:52:56 ◼   ► like I have to do, of course, for writing my review,

00:52:59 ◼   ► if that's something that you do,

00:53:02 ◼   ► whether you're a developer, whether you're writing a review,

00:53:04 ◼   ► or whatever you're doing with the betas,

00:53:06 ◼   ► if you're one of the people who has an account

00:53:09 ◼   ► like for a reason now because you're part of the press

00:53:12 ◼   ► or because you're a developer,

00:53:13 ◼   ► you almost never use it against your real data.

00:53:16 ◼   ► So you're not gonna use your real email account

00:53:19 ◼   ► or your real contacts, and with good reason,

00:53:21 ◼   ► because if you had done that,

00:53:22 ◼   ► you would learn eventually that some early beta version

00:53:24 ◼   ► of an OS scrambled all your contacts and now you're sad.

00:53:27 ◼   ► You always have usually multiple test accounts,

00:53:29 ◼   ► multiple test iCloud accounts,

00:53:31 ◼   ► multiple game center accounts, multiple everything.

00:53:33 ◼   ► Like, you know, dummy email accounts, dummy contacts,

00:53:37 ◼   ► like you don't put it on your real data.

00:53:39 ◼   ► And so are you gonna notice that the Gmail bug,

00:53:42 ◼   ► if all you're doing is just like,

00:53:44 ◼   ► you're not actually sending and receiving email,

00:53:46 ◼   ► you're not actually keeping track

00:53:47 ◼   ► of what you're marking red and what you haven't.

00:53:48 ◼   ► And so you wouldn't notice these bugs

00:53:50 ◼   ► because you're like, yeah, yeah, yeah, mail.

00:53:52 ◼   ► Like why would you even go to the mail app

00:53:53 ◼   ► unless you just wanted to see

00:53:54 ◼   ► what the new features are, whatever.

00:53:55 ◼   ► If you don't use it for your real email every day,

00:53:57 ◼   ► like, huh, didn't I mark that messages on Reddit?

00:53:59 ◼   ► It's just a bunch of test messages.

00:54:00 ◼   ► Maybe you wouldn't notice at all.

00:54:02 ◼   ► And I think this actually will go to a different pool

00:54:05 ◼   ► of people, pool of enthusiasts who want to use it

00:54:07 ◼   ► and who are going to use it with their real data.

00:54:09 ◼   ► And I would say at this point,

00:54:11 ◼   ► if you find this program exciting

00:54:12 ◼   ► and you wanna use these betas,

00:54:14 ◼   ► keep in mind that like Marco said,

00:54:15 ◼   ► he doesn't use it on his main machine.

00:54:17 ◼   ► I never installed beta versions of OS X on my machine ever.

00:54:21 ◼   ► Like I have complete separation because I,

00:54:24 ◼   ► and even when it's like, oh, this is the golden master,

00:54:26 ◼   ► like, oh, have you installed the ultimate?

00:54:28 ◼   ► No, I wait until I can get it from these days from the Mac App Store.

00:54:32 ◼   ► And didn't they have a thing this most recent release where like the build number changed

00:54:36 ◼   ► at the last minute or whatever?

00:54:37 ◼   ► I never ever ever install anything from ADC on my real computer.

00:54:41 ◼   ► I always get it, I used to, you know, I would always get it from the actual CD that would

00:54:44 ◼   ► come in the mail, you know, like a retail CD or DVD, or these days the actual one from

00:54:50 ◼   ► the Mac App Store.

00:54:51 ◼   ► I won't even usually install the .0 release.

00:54:53 ◼   ► Yeah, well, occasionally I'll be worried about that,

00:54:57 ◼   ► but in the past few releases they've been pretty good.

00:54:59 ◼   ► So if you're thinking of doing this,

00:55:01 ◼   ► Apple has all these crazy warnings and everything

00:55:02 ◼   ► and have it backed up or whatever.

00:55:04 ◼   ► By all means, if you're an enthusiast

00:55:06 ◼   ► and you know what you're getting into,

00:55:07 ◼   ► but realize it could erase your entire hard drive.

00:55:10 ◼   ► It could scramble all your contacts in iCloud.

00:55:11 ◼   ► You're like, "Oh, don't worry, I got backups."

00:55:12 ◼   ► What if it scrambles all your contacts in iCloud?

00:55:14 ◼   ► Where are those backups?

00:55:15 ◼   ► How are you gonna restore that?

00:55:16 ◼   ► iCloud is very opaque

00:55:17 ◼   ► and it's difficult to wrangle it back into shape.

00:55:19 ◼   ► So be aware of what you're getting into.

00:55:21 ◼   ► I imagine that since this is free,

00:55:23 ◼   ► A lot of people are going to be like, oh, awesome, I'm going to try the beta.

00:55:25 ◼   ► And then it will hose them in some way.

00:55:26 ◼   ► They'll be like, wah.

00:55:27 ◼   ► The beta hose.

00:55:28 ◼   ► It's like, that's, it's, you know, that's, you knew what you were getting into,

00:55:32 ◼   ► but you don't know, like your enthusiasm kind of outrun.

00:55:34 ◼   ► And that's part of the experience.

00:55:35 ◼   ► I did the same thing back in the day with like, you know, Mecha was a betas and

00:55:39 ◼   ► stuff where they were just terrible and data destroying.

00:55:42 ◼   ► And it's, that's, I guess it's okay.

00:55:46 ◼   ► I'm trying not to generalize and say it's going to be like a bunch of enthusiastic

00:55:49 ◼   ► teenagers destroying their parents' Macs with beta versions of it.

00:55:52 ◼   ► but like I was that teenager, you know, why not? Why shouldn't it be? Like, I think that's

00:55:57 ◼   ► kind of a rite of passage. But, so, basically what I'm saying is I would recommend nobody

00:56:01 ◼   ► join this program, but I think that people will. And I think that the people who do join

00:56:06 ◼   ► the program will be providing a different kind of testing than the existing people who

00:56:10 ◼   ► pay $99 a year.

00:56:13 ◼   ► We are also sponsored this week. I feel good about the sponsor reads this week because

00:56:16 ◼   ► they printed on card stock. See?

00:56:17 ◼   ► Oh, thanks, Merlin.

00:56:19 ◼   ► because of a printer loading error.

00:56:22 ◼   ► So I have nice thick sponsor reads this week.

00:56:25 ◼   ► - PC load letter.

00:56:26 ◼   ► - Exactly, what does that mean?

00:56:27 ◼   ► - Office space.

00:56:29 ◼   ► - Hey, all right.

00:56:30 ◼   ► - Is that both of us?

00:56:32 ◼   ► - Yeah.

00:56:32 ◼   ► - All right.

00:56:34 ◼   ► It's first time for everything kids.

00:56:36 ◼   ► - All right, New Relic is an all-in-one web application

00:56:39 ◼   ► performance management tool, APM.

00:56:42 ◼   ► It lets you see performance from the end user experience

00:56:44 ◼   ► down through your servers and down to each line

00:56:46 ◼   ► of your server side code.

00:56:48 ◼   ► Our friends at New Relic asked us to take a minute and say a big thank you to all your data nerds out there building all this great stuff that we all know and love.

00:56:56 ◼   ► They're sending a shout out to the developers, the software geeks, the code jockeys, to those brave few who see things differently.

00:57:03 ◼   ► High fives to all you rule breakers and disruptors.

00:57:07 ◼   ► Here's to working nights, to wearing oversized concentration-enhancing headphones upon your furrowed brows.

00:57:12 ◼   ► New Relic thanks you.

00:57:14 ◼   ► Nowadays, if you're in any business, you're in the software business.

00:57:18 ◼   ► Software powers our apps, runs our databases, manages our accounts,

00:57:21 ◼   ► and runs e-commerce sites and email programs. When software breaks,

00:57:25 ◼   ► everyone loses. New Relic helps improve your software performance

00:57:29 ◼   ► so your users have a better experience and your business is more successful.

00:57:32 ◼   ► How's that for a win-win? New Relic monitors

00:57:36 ◼   ► every move your application makes across the entire stack and shows you what's

00:57:39 ◼   ► happening

00:57:40 ◼   ► right now. You can zero in on problems quickly with transaction tracing, SQL, and NOSQL

00:57:46 ◼   ► performance analytics. I'm going to go with that now, I think. Application topography

00:57:51 ◼   ► mapping and deployment history makers and comparisons. See, I don't even know what half

00:57:56 ◼   ► these things are, but they sound awesome. If I was actually up to date with web development,

00:58:00 ◼   ► I would probably know what these things are, but I don't. But this sounds really awesome.

00:58:04 ◼   ► They even support PHP.

00:58:05 ◼   ► So they support Ruby, PHP, Java, .NET, Python,

00:58:08 ◼   ► and even Node.

00:58:10 ◼   ► So if you're one of those crazy cutting edge Node developers

00:58:12 ◼   ► like Casey, you can actually use New Relic right now

00:58:16 ◼   ► to monitor your application's performance.

00:58:17 ◼   ► So go to newrelic.com/atp for a 30 day free trial.

00:58:22 ◼   ► All you gotta do is deploy their agent

00:58:24 ◼   ► and it'll start collecting data immediately

00:58:26 ◼   ► and you'll be able to quickly see inside your app

00:58:28 ◼   ► to start finding hotspots, fixing issues,

00:58:30 ◼   ► and optimizing performance.

00:58:32 ◼   ► Once again, go to newrelic.com/ATP for a free 30-day trial.

00:58:36 ◼   ► Thanks a lot to New Relic for sponsoring our show once again.

00:58:41 ◼   ► We learned today about some new Thunderbolt information,

00:58:45 ◼   ► about third generation Thunderbolt.

00:58:46 ◼   ► And I'm not really that excited about it,

00:58:50 ◼   ► mostly because I don't plan on getting a new computer anytime

00:58:53 ◼   ► soon.

00:58:53 ◼   ► But Marco, do you want to tell us about this?

00:58:57 ◼   ► Well, there's not a whole lot known yet.

00:59:00 ◼   ► So there's not really much to discuss.

00:59:02 ◼   ► I might as well just save it for three years from now

00:59:05 ◼   ► when these ports actually come out.

00:59:06 ◼   ► But basically, Intel,

00:59:09 ◼   ► there was some leaked Intel roadmap document

00:59:12 ◼   ► that showed details on the next Thunderbolt version,

00:59:15 ◼   ► Thunderbolt 3, it'll probably be called.

00:59:17 ◼   ► And basically, we'll put the link in the show notes

00:59:20 ◼   ► on MacRumors, basically doubling the bandwidth,

00:59:23 ◼   ► so up to 40 gigabits per second.

00:59:25 ◼   ► And they're gonna actually change the connector,

00:59:29 ◼   ► which is probably gonna be interesting in the community.

00:59:31 ◼   ► They're gonna make a smaller connector

00:59:34 ◼   ► that also apparently will be able to charge

00:59:36 ◼   ► up to 100 watts.

00:59:38 ◼   ► - Yeah, that's ridiculous.

00:59:40 ◼   ► - Underscore Kyle Cronin in the chat room

00:59:42 ◼   ► just asked a really good question.

00:59:43 ◼   ► Which direction does the power go here?

00:59:46 ◼   ► Does the peripheral power the computer or vice versa?

00:59:49 ◼   ► - I thought you asked that on Twitter already.

00:59:50 ◼   ► You were saying you could almost power a display.

00:59:52 ◼   ► Didn't you tweet that today?

00:59:53 ◼   ► - Yeah, that's interesting though that,

00:59:55 ◼   ► yeah, I didn't think about it that direction.

00:59:57 ◼   ► All this says is, enables system charging up to 100 watts.

01:00:02 ◼   ► Yeah, that actually--

01:00:02 ◼   ► ADC did it already.

01:00:03 ◼   ► Speaking of ADC, the other ADC, Apple Display Connector,

01:00:06 ◼   ► powered the 22-inch Apple Center of Display.

01:00:09 ◼   ► There was one cabling in it.

01:00:10 ◼   ► No, this is the other way.

01:00:12 ◼   ► This is the other way, though.

01:00:13 ◼   ► This is saying that a Thunderbolt display could

01:00:17 ◼   ► charge your laptop.

01:00:18 ◼   ► Right, without the little dangly MagSafe thing

01:00:20 ◼   ► like we have now.

01:00:21 ◼   ► Somebody pointed out on Twitter also

01:00:23 ◼   ► that this is the proposed,

01:00:25 ◼   ► the code name Alpine Ridge Controller from Intel.

01:00:28 ◼   ► And this is apparently like still three generations away

01:00:30 ◼   ► or something, it's pretty far off.

01:00:32 ◼   ► So we're probably not seeing it soon.

01:00:34 ◼   ► It's probably a year or two at least.

01:00:37 ◼   ► - I think it might be sooner than we think

01:00:39 ◼   ► because like it's maybe not across the line,

01:00:43 ◼   ► but soon enough that I think the change in connector

01:00:46 ◼   ► is a little bit upsetting 'cause it's like,

01:00:48 ◼   ► I know we didn't want Apple to wait any longer

01:00:50 ◼   ► to do what it's,

01:00:52 ◼   ► really want it to be longer for the Mac Pro. But even if it's like a one and a half to

01:00:57 ◼   ► three year gap, like you're going to change a Thunderbolt connector already? We just got

01:01:00 ◼   ► Thunderbolt ports on our Mac Pro. And then you're going to change them? And the difference

01:01:04 ◼   ► is 1.5 millimeters in height.

01:01:08 ◼   ► And I mentioned on Twitter that I had been staring at the edges of various Retina MacBook

01:01:12 ◼   ► Pros and meetings at work, because that's what you do in boring meetings, thinking about

01:01:17 ◼   ► which one of those connectors is going to get thinner next. Because if you look at the

01:01:20 ◼   ► the sides, what we have is kind of a menagerie of MacBook and MacBook Pros around the office,

01:01:26 ◼   ► depending on who got their hardware when. So you can see them slowly getting skinnier

01:01:29 ◼   ► and you can see like, oh, the ethernet port disappears, oh, the MagSafe gets squished,

01:01:33 ◼   ► oh, now it looks like the USB is practically edge to edge, and you can see them squeezing

01:01:36 ◼   ► and squeezing. And really, there's headroom there, but the way Apple designs their laptops,

01:01:41 ◼   ► there is a curved section that sort of gives you a place for you to hook your fingers underneath,

01:01:45 ◼   ► and then there's a flat section that's perpendicular to the table, and that's where the ports go.

01:01:50 ◼   ► And they squeeze that thing down.

01:01:52 ◼   ► MagSafe, USB, and Thunderbolt are squeezed pretty tightly in the flat section now.

01:01:57 ◼   ► If you want to make the laptop thinner, you could start giving up the little lip that

01:02:01 ◼   ► you tuck your fingers under, but you don't want to make something that's as thin as just

01:02:04 ◼   ► a flat section and flat on a table.

01:02:05 ◼   ► You wouldn't be able to pick it up.

01:02:06 ◼   ► You'd be trying to get your fingernails underneath it to try to pull it up off the table.

01:02:10 ◼   ► So things have to get thinner.

01:02:12 ◼   ► And Thunderbolt is probably, USB is probably the first one, but we already know, we talked

01:02:16 ◼   ► about it in the past show, the new USB 3 connector that it'll be skinnier.

01:02:19 ◼   ► so that takes care of the USB ports.

01:02:22 ◼   ► That leaves Thunderbolt as potentially the thickest thing

01:02:24 ◼   ► besides MagSafe, and MagSafe I think is due

01:02:26 ◼   ► to be revised as well because MagSafe 2,

01:02:29 ◼   ► everyone seems to hate, but due to the design,

01:02:32 ◼   ► it's like, well, it's a big magnetic surface,

01:02:34 ◼   ► and people hate MagSafe 2 because it's not as secure

01:02:36 ◼   ► as the old one because they made it skinnier.

01:02:39 ◼   ► If you make it skinnier still, won't it be even worse,

01:02:41 ◼   ► and maybe it's time for a new magnetic power connector

01:02:44 ◼   ► that, I mean, I don't think they want it to insert,

01:02:46 ◼   ► but the whole point, it's supposed to come off easily,

01:02:48 ◼   ► but I think they could use a designer revision as well.

01:02:51 ◼   ► But Thunderbolt, it seems kind of early for them

01:02:53 ◼   ► to be making the thing thinner,

01:02:55 ◼   ► and I'm a little bit worried about that.

01:02:56 ◼   ► I hope this bus comes sooner rather than later.

01:02:59 ◼   ► I hope actually 2015 that we would see this,

01:03:01 ◼   ► sometime in calendar year 2015,

01:03:03 ◼   ► if only for the retina angle,

01:03:06 ◼   ► because now we finally, in theory,

01:03:08 ◼   ► would have bandwidth to do my desired

01:03:11 ◼   ► quad 27-inch retina resolution display.

01:03:15 ◼   ► It'd still depend on the new display port specs

01:03:17 ◼   ► and everything else,

01:03:18 ◼   ► This lists a bunch of acronyms that could, in theory, provide this.

01:03:20 ◼   ► Like, I don't know what HDMI 2 encompasses.

01:03:22 ◼   ► I don't even know what TBT is.

01:03:25 ◼   ► So maybe one of those things could power a monitor at the resolution that I want.

01:03:31 ◼   ► But changing the connectors.

01:03:33 ◼   ► They mentioned there'd be an adapter.

01:03:34 ◼   ► An adapter?

01:03:35 ◼   ► You're going to get—it's like two years after Thunderbolt appears, and now we're

01:03:38 ◼   ► already at those adapters.

01:03:39 ◼   ► I guess I guess it's MagSafe 1 and 2 all over again.

01:03:41 ◼   ► Well, and to be fair, I mean, how many people really have Thunderbolt devices so far?

01:03:47 ◼   ► It's not going to be a massive problem to have all these people buy little $30 adapters

01:03:52 ◼   ► or probably even less than that.

01:03:56 ◼   ► The whole problem with Thunderbolt is that no one's using it.

01:03:58 ◼   ► Yeah, the stagnation of the market is an advantage because it's like, "Well, if I had anything

01:04:02 ◼   ► to plug in, I wouldn't need an adapter, but since I have nothing to plug in anyway..."

01:04:05 ◼   ► I guess even Apple's monitors, even for the people that are doing what I do with the MacBook

01:04:09 ◼   ► Air and the Thunderbolt display.

01:04:11 ◼   ► I bet a lot of people do that setup where you have Thunderbolt plugged into your laptop

01:04:16 ◼   ► So you don't have, it's the equivalent,

01:04:18 ◼   ► so we don't have to have a dock, you know,

01:04:19 ◼   ► the stupid docks, they have a PC laptop,

01:04:21 ◼   ► so you snap the thing in and 800 connectors connect.

01:04:23 ◼   ► We can do it all through Thunderbolt, right?

01:04:25 ◼   ► Well, now you're gonna need, like,

01:04:27 ◼   ► the poor Thunderbolt display.

01:04:28 ◼   ► First it came with the MagSafe one to two adapter,

01:04:30 ◼   ► like you used to get that with a display,

01:04:32 ◼   ► so you could, you know,

01:04:33 ◼   ► attach the charging thingy to the laptops,

01:04:35 ◼   ► and now it will also come with a Thunderbolt

01:04:37 ◼   ► one to two adapter.

01:04:39 ◼   ► (laughing)

01:04:40 ◼   ► - For what it's worth though,

01:04:41 ◼   ► ever since the Thunderbolt display came out,

01:04:43 ◼   ► I have lusted after it,

01:04:44 ◼   ► and I'm way too cheap to buy one,

01:04:46 ◼   ► but the thought of not having to have,

01:04:49 ◼   ► I don't know, four or five or six cables to plug in

01:04:52 ◼   ► every time I plug in my computer when I get home

01:04:54 ◼   ► and only have two just sounds awesome.

01:04:57 ◼   ► And to bring this down to just one,

01:04:59 ◼   ► if we're reading this right and the power would flow

01:05:03 ◼   ► from display into laptop, I mean, that sounds great.

01:05:06 ◼   ► It genuinely, genuinely does.

01:05:07 ◼   ► - The problem is that one connector's

01:05:09 ◼   ► probably gonna be crappy, 'cause I can tell you,

01:05:11 ◼   ► we don't disconnect our laptop a lot,

01:05:13 ◼   ► but I've done it enough that every time I do it,

01:05:15 ◼   ► it does not feel like a robust,

01:05:17 ◼   ► oh, you can connect and disconnect this 8,000 times

01:05:19 ◼   ► and it'll be fine every time I do it.

01:05:20 ◼   ► I'm so conscious of the potential number of connects

01:05:24 ◼   ► and disconnects this cable has in it

01:05:25 ◼   ► because there's a little chip in that thing

01:05:26 ◼   ► and it's like a long stiff thing

01:05:28 ◼   ► sticking out of the side of your laptop.

01:05:30 ◼   ► It's not as, you feel like I can plug and unplug

01:05:33 ◼   ► the lightning port in my iPhone forever

01:05:35 ◼   ► and I feel fine about it until maybe the cable will fray

01:05:37 ◼   ► but whatever, I'll get a new one.

01:05:39 ◼   ► The Thunderbolt thing going into the side

01:05:40 ◼   ► especially MacBook Air, feels pretty perilous,

01:05:44 ◼   ► and I have little faith that the 1.5 millimeter

01:05:46 ◼   ► thinner version will feel any less perilous,

01:05:48 ◼   ► unless they invert it and do like a lightning-style

01:05:50 ◼   ► connector, but I'm not holding my breath.

01:05:52 ◼   ► - Well, I've read somewhere, maybe this is just

01:05:54 ◼   ► somebody's reaction to this, but I did read somewhere

01:05:56 ◼   ► that maybe one of the reasons they were redesigning

01:05:58 ◼   ► the connector is that the new one really isn't that great,

01:06:01 ◼   ► that it does fall out easily,

01:06:03 ◼   ► or it becomes slightly pulled easily,

01:06:04 ◼   ► and there's really, like there's no part of its design

01:06:08 ◼   ► that's really high friction, or no little locking

01:06:10 ◼   ► ball bearings like the Lightning has or anything like that.

01:06:12 ◼   ► So it's, I've never felt that Thunderbolt

01:06:16 ◼   ► things were secure.

01:06:17 ◼   ► I mean, I mostly just use it like for ethernet.

01:06:19 ◼   ► - Yeah, that feels precarious, right?

01:06:21 ◼   ► - But yeah, exactly, exactly.

01:06:23 ◼   ► - All right, this is a little bit random,

01:06:26 ◼   ► but I keep meaning to ask you when we were asked on Twitter,

01:06:30 ◼   ► what's the Overcast update?

01:06:31 ◼   ► - I'm almost in beta.

01:06:35 ◼   ► That's basically it,

01:06:36 ◼   ► which is probably what I said last time,

01:06:37 ◼   ► but I'm way closer this time.

01:06:39 ◼   ► Today I was writing the directory to give you some idea,

01:06:42 ◼   ► making the ad podcast thing work more than just search.

01:06:46 ◼   ► That's what I'm doing right now.

01:06:49 ◼   ► It's getting close, the login thing is done,

01:06:52 ◼   ► all the account stuff is done.

01:06:54 ◼   ► Yeah, getting there, that's about it.

01:06:58 ◼   ► It isn't that interesting right now.

01:06:59 ◼   ► - That's fair, I just don't think we talked about it

01:07:02 ◼   ► in a while, that's quick and easy.

01:07:05 ◼   ► Are iPad sales leveling off?

01:07:08 ◼   ► There's been a little bit of chatter about this lately.

01:07:10 ◼   ► - Yeah, this I wanted to talk about.

01:07:12 ◼   ► So there was this post by, oh boy.

01:07:17 ◼   ► - Go ahead, John, save Marco.

01:07:20 ◼   ► - Please.

01:07:21 ◼   ► - Jean-Louis Gasset.

01:07:22 ◼   ► - Thank you so much.

01:07:23 ◼   ► All right, so he wrote this post,

01:07:27 ◼   ► which we'll link to in the show notes,

01:07:29 ◼   ► basically saying or suggesting that iPad sales

01:07:32 ◼   ► have or are about to be revealed to have leveled off.

01:07:36 ◼   ► And then Gruber linked to it

01:07:37 ◼   ► and added his own commentary as well.

01:07:40 ◼   ► And so the theory here is that maybe iPad

01:07:43 ◼   ► and overall tablet sales are not doing as,

01:07:46 ◼   ► they're not growing as well as everyone expected.

01:07:49 ◼   ► And maybe tablets have been overestimated

01:07:54 ◼   ► and maybe the expectations of tablets have been too high.

01:07:58 ◼   ► So I wanted to talk about this

01:08:00 ◼   ► because I personally have really had trouble

01:08:05 ◼   ► making my iPad useful to me.

01:08:07 ◼   ► And I've heard so many people who that's not the case for.

01:08:11 ◼   ► So many people are finding incredibly interesting,

01:08:14 ◼   ► creative ways to use their iPad to get all their work done

01:08:16 ◼   ► or just bring their iPad when they go on trips

01:08:18 ◼   ► and not bring a laptop or whatever else.

01:08:20 ◼   ► And I have not had that experience at all.

01:08:24 ◼   ► I mean, and part of that's just 'cause the work I do,

01:08:26 ◼   ► you know, being a lot of programming and stuff like that.

01:08:28 ◼   ► But even things that aren't programming,

01:08:30 ◼   ► I find myself, when I'm on the iPad,

01:08:33 ◼   ► I find myself like saving things

01:08:35 ◼   ► for when I go back to my Mac because it'll just be so much faster and easier to do it

01:08:38 ◼   ► there.

01:08:39 ◼   ► You know, for me, the limitations of iOS work on a phone. The limitations of how apps are

01:08:46 ◼   ► so incredibly isolated from each other and so multitasking is very limited. There is

01:08:52 ◼   ► no like, you know, the document model of iOS where you don't have your files, you just

01:08:58 ◼   ► have each app and it has its own little silo of data. You know, all that stuff makes it

01:09:03 ◼   ► very hard for me to work the way I want to work.

01:09:07 ◼   ► And so I have not found the iPad to be very useful,

01:09:09 ◼   ► and in fact I hardly ever use mine.

01:09:11 ◼   ► I have stopped buying every version of it,

01:09:14 ◼   ► even for developer testing purposes.

01:09:17 ◼   ► Now that I have the Retina Mini,

01:09:19 ◼   ► and my wife has the iPad Air,

01:09:21 ◼   ► I really can't see us buying the next ones

01:09:23 ◼   ► that come out either,

01:09:24 ◼   ► 'cause these are gonna be fine for a while.

01:09:26 ◼   ► - Yeah, I'll remind you of that.

01:09:27 ◼   ► - Sounds like you need an iPad Pro, Margot.

01:09:30 ◼   ► (laughing)

01:09:31 ◼   ► And so I wonder, how many people is this the case for?

01:09:35 ◼   ► So for me, what I've found,

01:09:37 ◼   ► and this is why I think the iPhone lineup

01:09:40 ◼   ► might make this more interesting,

01:09:42 ◼   ► I've found that there's things that a phone does well,

01:09:47 ◼   ► and the iPad does most of those things well,

01:09:49 ◼   ► but the things that the phone does poorly,

01:09:52 ◼   ► the iPad doesn't do a lot of them much better.

01:09:56 ◼   ► And the iPad is worse than the phone in a few major ways.

01:09:59 ◼   ► Number one, it is not always with you.

01:10:01 ◼   ► Your phone is always in your pocket.

01:10:03 ◼   ► Your iPad, you might not have it with you.

01:10:05 ◼   ► In my case, I almost never have it with me.

01:10:08 ◼   ► 'Cause it doesn't fit in clothing or jacket pockets

01:10:11 ◼   ► and I don't carry bags big enough for it

01:10:12 ◼   ► most of those places I go.

01:10:13 ◼   ► So, you know, the phone is always with you

01:10:17 ◼   ► and your phone always has a data connection for most people.

01:10:20 ◼   ► iPads, yes, you can get them with data connections

01:10:22 ◼   ► but it's not the common case, I don't think.

01:10:26 ◼   ► And even if you have data connection,

01:10:27 ◼   ► like I feel kind of bad paying 20 bucks a month

01:10:31 ◼   ► for this data connection that I hardly ever use,

01:10:33 ◼   ► but my iPad becomes even less useful if I don't have that.

01:10:36 ◼   ► But I'm not happy about that.

01:10:38 ◼   ► Most people just don't get data connections

01:10:40 ◼   ► or data plans for their iPads.

01:10:42 ◼   ► And so, you know, that major difference

01:10:45 ◼   ► between always being with you

01:10:47 ◼   ► and always having a data connection,

01:10:50 ◼   ► that actually makes the iPad even less useful for me

01:10:53 ◼   ► than it normally is.

01:10:55 ◼   ► And so I have to wonder, like, you know, I've said before,

01:11:00 ◼   ► I don't think it's wise to ever bet against the smartphone.

01:11:03 ◼   ► I think the smartphone is an amazing ideal

01:11:08 ◼   ► and it's like the intersection of the best of so many things

01:11:11 ◼   ► and it's so good and it's so useful

01:11:13 ◼   ► and there's so much potential in the smartphone.

01:11:16 ◼   ► For me, if the smartphone had a little bit bigger screen,

01:11:19 ◼   ► which it's probably about to,

01:11:21 ◼   ► which in the non-Apple world, that's old news,

01:11:25 ◼   ► and they've had big screens for years,

01:11:28 ◼   ► I really don't think I'm gonna have a use for an iPad

01:11:32 ◼   ► at all.

01:11:34 ◼   ► Even today, if my iPad got stolen today,

01:11:37 ◼   ► please don't come steal it, but if it got stolen today,

01:11:40 ◼   ► I'm pretty sure I would not replace it.

01:11:43 ◼   ► And so, but you know, if my, like my phone,

01:11:47 ◼   ► the combination of a phone and a computer

01:11:49 ◼   ► is so great for me and I'm wondering, you know,

01:11:52 ◼   ► whether the iPad is really necessary long term

01:11:55 ◼   ► as a product category and whether it's succeeding

01:11:57 ◼   ► long-term as a product category, or whether it will become more of a narrow niche product

01:12:02 ◼   ► as most people realize that a phone and maybe a computer is good enough.

01:12:08 ◼   ► So I would like to argue with everything you just said.

01:12:12 ◼   ► Cool.

01:12:13 ◼   ► So I had an iPad One.

01:12:16 ◼   ► I had an iPad 3rd gen, so the first Retina iPad.

01:12:20 ◼   ► And now I have what I like to call the RetinaPad Mini, which drives Stephen Hackett nuts, which

01:12:25 ◼   ► which makes me want to call it that even more.

01:12:28 ◼   ► And I got the RetinaPad Mini with LTE for the first time.

01:12:32 ◼   ► I'd always had WiFi-only iPads,

01:12:34 ◼   ► and this one I got with LTE.

01:12:36 ◼   ► I got a Verizon RetinaPad Mini,

01:12:38 ◼   ► and then right around the time I got it,

01:12:41 ◼   ► T-Mobile started doing that crazy thing

01:12:42 ◼   ► where you can buy a SIM for $10 one time,

01:12:46 ◼   ► and then they'll give you 200 megs of data for free

01:12:49 ◼   ► every single month.

01:12:51 ◼   ► And the idea is, or I presume the idea is that

01:12:54 ◼   ► It's a gateway drug to get you to pay T-Mobile for data.

01:12:58 ◼   ► Well, what I found is the combination of moving from big iPad to iPad mini,

01:13:06 ◼   ► as paired with this T-Mobile plan that really, quite honestly, gives me enough

01:13:11 ◼   ► data for pretty much any usage I would want away

01:13:14 ◼   ► from a wi-fi hotspot like a friend's house or family

01:13:17 ◼   ► member's house or at home or work, the combination of having an LTE iPad

01:13:22 ◼   ► with just a little bit of data every single month and having it be an iPad Mini made me fall in love

01:13:30 ◼   ► with this iPad Mini so much more than I did my prior iPads, which I kind of went in and out of

01:13:37 ◼   ► love with. I would use them a lot, then I wouldn't use them at all. Then I would use them a lot,

01:13:40 ◼   ► not at all. And I love my iPad Mini. And I wonder if, I wonder if the, by virtue of it having LTE

01:13:52 ◼   ► And either if I could tether against my iPhone, which I can't because I'm still on the AT&T

01:13:57 ◼   ► Unlimited plan, or if I had a plan wherein I could share my data with my iPad for very

01:14:04 ◼   ► little extra per month.

01:14:06 ◼   ► Any of these combinations or the T-Mobile thing that I've done makes me absolutely

01:14:10 ◼   ► adore my iPad.

01:14:11 ◼   ► And generally speaking, when I travel, I don't bring a computer with me unless I'm traveling

01:14:18 ◼   ► for work.

01:14:19 ◼   ► And in fact, we were traveling all of last week.

01:14:23 ◼   ► And the only reason I had a computer with me

01:14:25 ◼   ► was because I had to record the podcast

01:14:27 ◼   ► while we were on the road.

01:14:28 ◼   ► And if it wasn't for that,

01:14:29 ◼   ► I would have left the computer at home.

01:14:31 ◼   ► So it's funny because everything you just said,

01:14:34 ◼   ► it makes sense to me, but I like my iPad

01:14:37 ◼   ► for all the reasons you don't like your iPad.

01:14:40 ◼   ► I know, John, what do you think about this?

01:14:43 ◼   ► - I think the thing that's holding the iPad back

01:14:45 ◼   ► is two aspects.

01:14:46 ◼   ► One is that, as I said in past shows when I was talking about the iPad Pro, if the iPad

01:14:53 ◼   ► wants to take over more of the functionality currently occupied by PCs, it has to expand

01:14:57 ◼   ► its ability to do the things that PCs can do in some vague, as yet unspecified way,

01:15:03 ◼   ► possibly having a larger screen, possibly better multitasking, possibly better document

01:15:08 ◼   ► management, better inter-application communication, all the things that we discussed, right?

01:15:11 ◼   ► So that is, it's like, okay, well, what can you do with that?

01:15:13 ◼   ► Well, the PC market is not as big as the phone market.

01:15:16 ◼   ► Phone market's growing.

01:15:17 ◼   ► PC market, I think, is either growing slowly

01:15:20 ◼   ► or possibly shrinking.

01:15:21 ◼   ► So say the iPad takes over some portion of that.

01:15:25 ◼   ► That still doesn't get you the kind of growth

01:15:26 ◼   ► that people are expecting from it.

01:15:28 ◼   ► It's still like gangbusters, but that's one area

01:15:31 ◼   ► where it probably has to move

01:15:33 ◼   ► if it wants to get some growth.

01:15:34 ◼   ► And the other area is it's gotta come way, way down in price

01:15:37 ◼   ► because what it's competing against is psychologically,

01:15:40 ◼   ► if not in actuality,

01:15:41 ◼   ► competing against cell phones which are free. And they're not free. Like, "Oh, they cost thousands

01:15:47 ◼   ► of dollars a year for your cell phone plans, and iPhone's incredibly expensive." All of that is

01:15:50 ◼   ► true. But psychologically speaking, cell phones are practically free, because most people in

01:15:56 ◼   ► first world countries feel like they need to have a cell phone, and eventually all cell phones will

01:15:59 ◼   ► be what we currently call smartphones. And it's like, you know, it's not free because it costs

01:16:05 ◼   ► a lot of money, but people's like, "Well, I have to have a phone." Right, like the tablet is

01:16:08 ◼   ► is competing against nothing. If you don't buy a tablet, you will spend zero, because

01:16:16 ◼   ► no one's going to say, "Oh, you know what? I'm going to replace my phone with a tablet."

01:16:19 ◼   ► Well, that's the phablet phenomenon, right? But yeah, what they're competing against is

01:16:25 ◼   ► they consider a phone as like, "I have to pay for that. That's a sunk cost, and that's

01:16:30 ◼   ► basically free. Now you're telling me I've got to pay more?" And the thing is, if they

01:16:34 ◼   ► have some task they need to do that is just not convenient or possible to do on even a

01:16:38 ◼   ► very large phone. Their choice is a PC or a tablet, and PCs are traditionally more expensive,

01:16:44 ◼   ► but tablets, especially Apple's tablets, are still pretty darn expensive. If Apple could,

01:16:49 ◼   ► say the Retina iPad Mini was $99, just throw out some fantasy scenario there, there's no

01:16:55 ◼   ► doubt that that would move the needle on sales. And even if you think, "Well, I'm not sure

01:17:00 ◼   ► how I'm going to use this, and I'm not sure how that would shake out," but I just think

01:17:05 ◼   ► The idea that every single home that has a smartphone would also have an iPad just hanging

01:17:10 ◼   ► around just to be that thing that, like the modern, the future world equivalent of a magazine.

01:17:15 ◼   ► I'm just going to sit on the couch and read a magazine.

01:17:17 ◼   ► Well, you're not going to read a magazine.

01:17:18 ◼   ► You're going to look at webpages on your iPad.

01:17:20 ◼   ► Or I'm just going to go up to my room and watch some Netflix or stream something from

01:17:24 ◼   ► my DVR or whatever.

01:17:25 ◼   ► Like the idea that just having a flat screen, even if it never leaves the house, it's just

01:17:29 ◼   ► a convenient thing to have around the house.

01:17:31 ◼   ► and I think the uses of it are fine,

01:17:33 ◼   ► but when it costs $700, you're like,

01:17:35 ◼   ► do I need one of those Future World things, $700 worth?

01:17:38 ◼   ► No, I don't, even $500.

01:17:40 ◼   ► And I think they'll sell a lot more than

01:17:43 ◼   ► if they go down in price, and I guess there's three things.

01:17:45 ◼   ► The final thing is that many people have mentioned,

01:17:48 ◼   ► they're just a little bit too,

01:17:50 ◼   ► they're just a little bit too good so far.

01:17:51 ◼   ► I know so many people who are using iPad 1s

01:17:53 ◼   ► and see no reason to upgrade.

01:17:55 ◼   ► That is the, again, they're competing against the phone

01:17:57 ◼   ► where there's many reasons to upgrade

01:17:58 ◼   ► because of the subsidies and all the other things

01:18:00 ◼   ► or whatever, like, it's tough.

01:18:02 ◼   ► First of all, they don't feel like they should,

01:18:03 ◼   ► 'cause they say, "Well, I spent $500 on an iPad,

01:18:05 ◼   ► "I'm keeping it for years."

01:18:06 ◼   ► And the second thing is, well,

01:18:08 ◼   ► I guess unless games stop working,

01:18:11 ◼   ► they don't care if they don't run iOS 7.

01:18:13 ◼   ► I guess they do kind of cycle down to the kids

01:18:15 ◼   ► and let them use it, but it's just,

01:18:17 ◼   ► they are incredibly long-lasting and, I mean,

01:18:20 ◼   ► surprisingly durable for a $500 piece of glass.

01:18:22 ◼   ► You would think they'd all be destroyed by now,

01:18:24 ◼   ► but I've seen so many go through the grubby paws of kids

01:18:27 ◼   ► and just, they survive.

01:18:29 ◼   ► In a home environment, those things survive

01:18:30 ◼   ► for a long period of time.

01:18:32 ◼   ► And so, if there were $99, you'd be like,

01:18:34 ◼   ► "Eh, every year or two I'll get a new one

01:18:35 ◼   ► "of these flat things."

01:18:36 ◼   ► 'Cause it's what we use to like, you know,

01:18:38 ◼   ► it's what we use to like read stuff

01:18:40 ◼   ► and read, you know, read web pages.

01:18:42 ◼   ► It's better than reading it on my phone.

01:18:43 ◼   ► I won't drain my phone's battery.

01:18:44 ◼   ► And I use it to watch a video

01:18:45 ◼   ► 'cause it's a little bit bigger screen.

01:18:47 ◼   ► And if you could get the price way down,

01:18:49 ◼   ► and if you could get a little bit more capable,

01:18:51 ◼   ► that would make sure that there's an iPad

01:18:53 ◼   ► in everybody's home.

01:18:54 ◼   ► And that would also make sure that when people go on trips

01:18:57 ◼   ► So maybe I don't need to buy a laptop and I can eat into the PC market by just saying

01:19:01 ◼   ► I can take this with me because I can do more or less everything I need to do.

01:19:04 ◼   ► And you would say, when do you reach your limit?

01:19:06 ◼   ► Can developers do everything they need to do?

01:19:08 ◼   ► Probably not, but can you keep going down?

01:19:10 ◼   ► Can people do regular word processing work?

01:19:12 ◼   ► That's still pretty annoying.

01:19:13 ◼   ► That's where the iPad Pro comes in.

01:19:14 ◼   ► So it's not like Apple needs to like, "Hey, you need to sell your iPads at a loss right

01:19:18 ◼   ► now."

01:19:19 ◼   ► But I think just the natural...

01:19:20 ◼   ► Take a look at the Apple TV.

01:19:22 ◼   ► It took a while for the Apple TV to get down to $99.

01:19:26 ◼   ► The iPad is going to take even longer because that screen is tough to really push down much

01:19:30 ◼   ► farther, but give it five, seven years.

01:19:34 ◼   ► When Apple has tablet things that span the same range as iPods do, I think there is a

01:19:40 ◼   ► legitimate reason for every house that has a smartphone to have a tablet device.

01:19:45 ◼   ► If Apple can get its price point down, it could be that device.

01:19:47 ◼   ► I think that's where the growth potential is for tablets.

01:19:49 ◼   ► Yeah, that makes sense.

01:19:52 ◼   ► The thing that I think about is what is really holding the iPad especially, but the iPhone

01:20:00 ◼   ► and just iOS in general, what's holding it back?

01:20:04 ◼   ► And we've talked in the past around this time last year about what's the low-hanging

01:20:09 ◼   ► fruit for iOS.

01:20:12 ◼   ► And I guess maybe low-hanging fruit is a poor way of phrasing it, but what's the thing

01:20:18 ◼   ► that bothers everyone the most?

01:20:20 ◼   ► And copy-paste is the seminal example from years back, that it infuriated everyone to

01:20:27 ◼   ► not have copy-paste. Obviously, that had to show up quickly. Well, inter-app communication,

01:20:32 ◼   ► I feel like we're getting to the point that we're really going to need to crap or get

01:20:35 ◼   ► off the pot, or Apple is really going to need to. I mean, is this the—I don't want to

01:20:39 ◼   ► get into a WWDC prediction episode, but is this the time? Is that enough to make iOS

01:20:46 ◼   ► a little bit more powerful?

01:20:49 ◼   ► Interapp communication itself is a phrase we throw around, it's a concept we throw around

01:20:55 ◼   ► that we think, first of all, we all assume that iOS 8 will improve this somehow because

01:21:02 ◼   ► there's a lot of, you can see that there's a lot of the groundwork already laid in 6

01:21:05 ◼   ► and 7 for this. You have things like remote view controllers, things like the UI activity

01:21:09 ◼   ► system. You can tell that they're taking steps towards allowing this. But the problem

01:21:16 ◼   ► is that Internet communication itself does not really solve the problem. It makes some

01:21:24 ◼   ► things a little bit better. Suppose iOS does something very similar to the contracts you're

01:21:30 ◼   ► intents on Windows Phone and Android, which is what most developers are asking them to

01:21:35 ◼   ► do, myself included. I think that would be great. But I think a big part of the problem

01:21:40 ◼   ► is the document model, the file system model.

01:21:44 ◼   ► You know, where you store your data, how you access your stored data,

01:21:49 ◼   ► where things are

01:21:52 ◼   ► divided, and where things are not accessible from.

01:21:56 ◼   ► And I think that, you know, there's

01:21:59 ◼   ► all this simplicity on these devices of not having to manage your files,

01:22:04 ◼   ► not having to figure out where things are, not having to

01:22:06 ◼   ► go to your parents' house and see everything's on the desktop and that their

01:22:09 ◼   ► hard drive is "full" because they're out of space for their icons.

01:22:14 ◼   ► You avoid all those problems, but the amount of power that that removes from usage is pretty

01:22:20 ◼   ► dramatic.

01:22:21 ◼   ► And I think that, more than anything, is what limits iOS from being used in more productivity-type

01:22:27 ◼   ► roles.

01:22:28 ◼   ► Yeah, Apple's still holding the line on that, and I think most people agree with the sentiment,

01:22:36 ◼   ► And we just hope that, like, well, presumably Apple

01:22:40 ◼   ► has thought of this and has some solution.

01:22:42 ◼   ► Like, the geeks just say, oh, just show us the file system,

01:22:44 ◼   ► and they just want to reproduce all the same problems.

01:22:46 ◼   ► And like, yeah, that would solve it for geeks, right?

01:22:48 ◼   ► But-- modular sandboxing, I guess.

01:22:50 ◼   ► But for everybody else, it's like, well, siloed per app

01:22:54 ◼   ► is simple, but it has all these problems.

01:22:56 ◼   ► And we're all hoping for a solution that does not

01:22:59 ◼   ► give up all the simplicity and go all the way back

01:23:02 ◼   ► to just raw access to the file system,

01:23:03 ◼   ► again, if only for security concerns.

01:23:06 ◼   ► but has the benefits of being able to like,

01:23:08 ◼   ► you know, work, even normal people seem to grasp Dropbox

01:23:13 ◼   ► because Dropbox essentially took the file system problem

01:23:15 ◼   ► and narrowed it down to a single place,

01:23:17 ◼   ► kind of like the desktop, which was the old place

01:23:19 ◼   ► that everyone, you mentioned your icon-swelling desktop.

01:23:21 ◼   ► That was the one place that people felt comfortable,

01:23:23 ◼   ► I just put everything in desktop.

01:23:24 ◼   ► And then what people would do is like,

01:23:25 ◼   ► well, I have all my crap on my desktop,

01:23:27 ◼   ► is this the one place I know how to find?

01:23:29 ◼   ► And I make folders on my desktop for subprojects.

01:23:31 ◼   ► And it's like training wheels for the real file system.

01:23:34 ◼   ► And Dropbox is like that as well.

01:23:35 ◼   ► and it's like, it's the same as the file system.

01:23:38 ◼   ► It's just narrowed,

01:23:39 ◼   ► like your Dropbox is equivalent to the desktop.

01:23:40 ◼   ► But once you start making folders in your Dropbox,

01:23:42 ◼   ► it's like the person who finally starts making folders

01:23:44 ◼   ► on the desktop because lots of stuff.

01:23:45 ◼   ► And they start organizing things.

01:23:46 ◼   ► And like, you just wanted to say,

01:23:47 ◼   ► just go a little bit farther.

01:23:48 ◼   ► Like the whole file system is your oyster.

01:23:50 ◼   ► It's not just the desktop.

01:23:52 ◼   ► - New folder one.

01:23:53 ◼   ► - Yeah, you can put folders anywhere.

01:23:54 ◼   ► - New folder 17.

01:23:55 ◼   ► - Yeah, so I hope that Apple doesn't give up,

01:24:01 ◼   ► but at a certain point, like,

01:24:04 ◼   ► it's kind of like that, you know,

01:24:05 ◼   ► that give up and use Table's CSS joke site from years ago.

01:24:08 ◼   ► It's like at a certain point, it's like, well,

01:24:09 ◼   ► Apple, you've been thinking about this for years,

01:24:11 ◼   ► and if you really can't think of something that's better,

01:24:13 ◼   ► at a certain point, you need to just like say,

01:24:16 ◼   ► we failed in our ability to give you something

01:24:18 ◼   ► that's both simpler and equally powerful

01:24:19 ◼   ► or close to as powerful,

01:24:21 ◼   ► and just give us back the freaking file system.

01:24:23 ◼   ► And I do not want them to do that.

01:24:24 ◼   ► I think that would be not only an admission of failure,

01:24:27 ◼   ► but an actual failure.

01:24:28 ◼   ► So I really keep hoping that they come up with some scheme

01:24:32 ◼   ► that has the security benefits of sandboxing,

01:24:36 ◼   ► but that lets applications easily work,

01:24:38 ◼   ► multiple applications work on the same document.

01:24:40 ◼   ► I mean, that's part of my whole vague,

01:24:41 ◼   ► cloudy iPad Pro vision,

01:24:43 ◼   ► is the idea of using multiple applications

01:24:45 ◼   ► to make a sort of composite document,

01:24:47 ◼   ► or work on a project that has some images,

01:24:50 ◼   ► and some text files, and some this, and some that,

01:24:52 ◼   ► like you would do on your Mac

01:24:54 ◼   ► when you're doing any sort of remotely complicated project,

01:24:57 ◼   ► without it being, you know,

01:25:00 ◼   ► oh, I just do everything in one app.

01:25:01 ◼   ► Like even Xcode's going that direction.

01:25:02 ◼   ► Xcode is like, yes, it's a whole bunch of different things.

01:25:05 ◼   ► You've got images, you've got header files,

01:25:08 ◼   ► you've got source code, you have interface builder documents,

01:25:11 ◼   ► you've got all this stuff, but like it's all in Xcode.

01:25:13 ◼   ► I guess some of the image editor,

01:25:14 ◼   ► once they embed an image editor in Xcode,

01:25:16 ◼   ► we'll know they've really just given up.

01:25:19 ◼   ► 'Cause then they can just port Xcode 7 right to the iPad Pro

01:25:22 ◼   ► and it's like, hey, you don't need to launch

01:25:23 ◼   ► a separate app, all your stuff is in one place.

01:25:26 ◼   ► - Does it seem like the iPad app library is stagnating?

01:25:31 ◼   ► Because it seems that way to me, but because I'm not that into using my iPad, that could

01:25:36 ◼   ► just be because I use Tweetbot and it hasn't been updated yet, or whatever.

01:25:41 ◼   ► What do you mean by that?

01:25:42 ◼   ► Which kinds of apps are you talking about, specifically?

01:25:44 ◼   ► It seems to me that, again, anecdotally, so whatever this is worth, I'm not seeing a lot

01:25:51 ◼   ► of iPad apps come out anymore that are interesting or exciting, and a lot of the previously good

01:25:58 ◼   ► apps still either have been updated to iOS 7 and are very dated or are just on really

01:26:06 ◼   ► slow update cycles so they're seemingly abandoned or some of them actually have been

01:26:09 ◼   ► abandoned, some of them have been updated but have been made worse. It seems like the

01:26:14 ◼   ► iPad was a priority for app developers for about two and a half years and now it's

01:26:20 ◼   ► just not. And I can point to lots of, if this is the case beyond just what I'm seeing,

01:26:26 ◼   ► I can think of lots of reasons why that might be the case.

01:26:30 ◼   ► I've been meaning to write a blog post about this,

01:26:33 ◼   ► but I keep forgetting to do it.

01:26:35 ◼   ► But one of the biggest would be the App Store pricing model

01:26:40 ◼   ► and the top list model and how that encourages

01:26:43 ◼   ► the race to the bottom in pricing

01:26:44 ◼   ► and very low prices made up at very large volume.

01:26:47 ◼   ► All of that depends on there being very large volume.

01:26:52 ◼   ► But if the iPad is not being used that much

01:26:54 ◼   ► to buy a bunch of apps, and if instead most of them

01:26:56 ◼   ► are being used to watch videos and browse the web,

01:26:59 ◼   ► and we know they don't sell nearly as well

01:27:01 ◼   ► as iPhones do already, maybe that's a big problem.

01:27:06 ◼   ► I mean, it certainly, it seems to me,

01:27:09 ◼   ► like maybe, like to me, one of the reasons

01:27:11 ◼   ► why I'm not using the iPad is because, like,

01:27:13 ◼   ► the Twitter app is better on my phone,

01:27:15 ◼   ► the RSS app is better on my phone,

01:27:17 ◼   ► like half the things I do on my iOS devices,

01:27:21 ◼   ► like the iPhone, the apps I have on my iPhone

01:27:24 ◼   ► are better than any I can find on the iPad.

01:27:26 ◼   ► Is this just me, or is this a broader thing?

01:27:29 ◼   ► - I think the sales numbers definitely aren't in floats

01:27:31 ◼   ► 'cause there's way more iPhones,

01:27:32 ◼   ► but I think there's a bifurcation of applications.

01:27:34 ◼   ► And in the beginning it was like,

01:27:36 ◼   ► well everyone's gonna make all their apps everywhere.

01:27:37 ◼   ► And then eventually people learned which apps

01:27:39 ◼   ► are better suited to which environment.

01:27:41 ◼   ► So to give an example, these days if a new game comes out

01:27:44 ◼   ► I don't have fear that there's not gonna be an iPad version.

01:27:47 ◼   ► And if there's anything that has to do

01:27:49 ◼   ► with reading experience, a magazine app, a comic reader,

01:27:54 ◼   ► something for watching video.

01:27:55 ◼   ► There's no concern that it's not gonna have an iPad version.

01:27:59 ◼   ► Or even anything for like a controlling application,

01:28:01 ◼   ► like the TiVo application.

01:28:02 ◼   ► I'm not concerned that's not gonna, you know,

01:28:04 ◼   ► but anything that's not sort of on the go.

01:28:06 ◼   ► 'Cause on the go style apps like Twitter clients

01:28:08 ◼   ► and stuff like that, I don't expect to find them on the iPad

01:28:11 ◼   ► 'cause those are like clearly focused on the phone.

01:28:12 ◼   ► Anything that is clearly something you'd wanna use

01:28:14 ◼   ► like a mobile type application,

01:28:17 ◼   ► that's gonna be phone first or phone only.

01:28:18 ◼   ► Whereas any application that benefits

01:28:21 ◼   ► from having a larger screen,

01:28:22 ◼   ► that's gonna be iPad only or iPad first.

01:28:25 ◼   ► And so I definitely do see that split.

01:28:27 ◼   ► And I have the same frustration and it's like,

01:28:30 ◼   ► well, I know most of the time I'm not reading Twitter

01:28:33 ◼   ► on my iPad, but I would like to have a universal version

01:28:36 ◼   ► of whatever my favorite Twitter client,

01:28:37 ◼   ► as a matter of fact I do 'cause I use Twitterrific,

01:28:39 ◼   ► but those type of applications like Glassboard, for example,

01:28:43 ◼   ► there's no iPad version of that and it frustrates me.

01:28:45 ◼   ► I kind of make sense, but on the other hand,

01:28:48 ◼   ► it's like a mobile on the go app, but it's like,

01:28:51 ◼   ► but I'm on my iPad, I have to run that 2X thing,

01:28:54 ◼   ► I just feel bad.

01:28:55 ◼   ► So I just think it's kind of the application splitting

01:28:57 ◼   ► and going to where they need to go.

01:28:58 ◼   ► And unfortunately, some of that split is,

01:29:00 ◼   ► if you need X number of sales to be profitable,

01:29:03 ◼   ► forget the iPad, because there's not that many iPads

01:29:05 ◼   ► out there, but I think a lot of it is just

01:29:07 ◼   ► appropriateness of screen size.

01:29:09 ◼   ► - Well, and to answer your question, Marco,

01:29:11 ◼   ► I'm gonna kind of ask myself and then answer it

01:29:14 ◼   ► with a different question, which is,

01:29:15 ◼   ► what have I been excited about?

01:29:17 ◼   ► What have I been excited about on the iPad lately?

01:29:19 ◼   ► And I'm really excited that the 1Password iOS 7 update just came out.

01:29:24 ◼   ► That's universal.

01:29:25 ◼   ► I'm waiting with bated breath for a Tweetbot update for iOS 7.

01:29:32 ◼   ► And Fantastical 2 came out, and I'm really pumped about that.

01:29:35 ◼   ► That's about it.

01:29:36 ◼   ► I mean, 3s came out for the iPad.

01:29:38 ◼   ► I was excited about that too.

01:29:39 ◼   ► But do you see what I'm saying?

01:29:41 ◼   ► To your point, there's not a lot that's been brand new that has happened on the iPad.

01:29:46 ◼   ► Fantastic Isle is a good example though that I've been really really really excited about

01:29:50 ◼   ► That's happened shoot months if not around a year now. Well, I mean there's games like if you're into playing games

01:29:57 ◼   ► I mean, there's a pretty pretty regular cycle of

01:30:00 ◼   ► Exciting new games that are available for the iPad that come out

01:30:04 ◼   ► I mean, it's not you know, like it's not there are tons of the games that come out

01:30:08 ◼   ► Most of them are crappy but I'm talking about the good games like the current one is like Monument Valley and before that was threes and

01:30:12 ◼   ► and you had things like The Room and Year Walk.

01:30:16 ◼   ► And there's always lots of great software out there,

01:30:19 ◼   ► games that take advantage of a large screen.

01:30:21 ◼   ► But I think a lot of the iPad stuff,

01:30:23 ◼   ► thinking of it as the thing you have around the house, yeah,

01:30:25 ◼   ► it's going to be a video app.

01:30:26 ◼   ► And through it, you're going to consume stuff.

01:30:28 ◼   ► Or it's going to be a comic reader through which you're

01:30:29 ◼   ► going to spend a continual stream of money on comics.

01:30:32 ◼   ► I mean, just ask Merlin.

01:30:33 ◼   ► Like, is he excited that there's a new version of Comixology

01:30:36 ◼   ► out?

01:30:36 ◼   ► No, but it's just like it's a fixture.

01:30:38 ◼   ► It becomes part of that appliance.

01:30:39 ◼   ► And it's an appliance through which you funnel money.

01:30:41 ◼   ► So it's not like there's not money to be made there,

01:30:43 ◼   ► because if you have that app and use it to buy things,

01:30:46 ◼   ► or even just magazine subscriptions,

01:30:48 ◼   ► it's a continual faucet of money,

01:30:51 ◼   ► versus just like, oh, I can sell you one app for $1.99

01:30:54 ◼   ► if I'm lucky.

01:30:55 ◼   ► It's like, no, you'll spend 50, $100 a month on comics

01:30:58 ◼   ► through this one app that we will continually update

01:31:00 ◼   ► and develop, but it's not exciting.

01:31:02 ◼   ► It's not like there's amazing new apps coming out.

01:31:04 ◼   ► But I think there's less of that on the phone,

01:31:06 ◼   ► with the exception possibly of music streaming services

01:31:09 ◼   ► that you pay a subscription for.

01:31:10 ◼   ► the phone is less of a venue for,

01:31:13 ◼   ► I'm gonna download this one app

01:31:14 ◼   ► and yeah, the app will be updated,

01:31:16 ◼   ► but for the next year or two years or my entire life,

01:31:18 ◼   ► I will continue to funnel money through this application

01:31:20 ◼   ► 'cause it's the way I get my video, my comics,

01:31:23 ◼   ► or my games or whatever.

01:31:25 ◼   ► - Yeah, I mean, I think a lot of it also depends on,

01:31:28 ◼   ► you know, for any, whatever app you're talking about,

01:31:30 ◼   ► like how hard is it, how much work is it

01:31:33 ◼   ► to maintain a separate iPad interface

01:31:35 ◼   ► or a separate iPad app entirely?

01:31:37 ◼   ► And for certain types of things like,

01:31:39 ◼   ► You know, video type apps, it's a lot less work

01:31:42 ◼   ► than something like a word processor would be.

01:31:45 ◼   ► You know, like games, it's, you know,

01:31:48 ◼   ► some work in the interface layout and stuff,

01:31:50 ◼   ► but for the most part, you're getting that fairly easily.

01:31:54 ◼   ► There's, you know, there's all these productivity apps.

01:31:58 ◼   ► That's, it's a lot more work.

01:31:59 ◼   ► Like, you know, I'm not even making Overcast

01:32:02 ◼   ► for iPad when I launch.

01:32:03 ◼   ► And, you know, maybe I'll add it later,

01:32:04 ◼   ► but I'm launching without any iPad interface,

01:32:06 ◼   ► even though I could just say in the P list,

01:32:09 ◼   ► yeah, I work on the iPad, sure,

01:32:11 ◼   ► and just let the interface scale itself up,

01:32:14 ◼   ► but it would look terrible.

01:32:15 ◼   ► I'd rather not do that, and so I'm not doing that.

01:32:18 ◼   ► I'd rather do it well or not at all.

01:32:22 ◼   ► So right now I'm gonna choose not at all,

01:32:23 ◼   ► and then maybe I'll address well later.

01:32:26 ◼   ► But I wonder for how many developers,

01:32:29 ◼   ► how many developers are gonna make that call and say,

01:32:31 ◼   ► you know what, it's probably not worth it.

01:32:33 ◼   ► Like almost every other podcast app on the store

01:32:36 ◼   ► that's popular at all has a dedicated iPad version.

01:32:39 ◼   ► 'Cause all of them were written like two to three years ago

01:32:41 ◼   ► when that was a thing you all did.

01:32:43 ◼   ► Like everyone, you had to have an iPad version

01:32:45 ◼   ► two or three years ago 'cause it was like the hotness.

01:32:48 ◼   ► But now, for so many types of apps,