12: Accidental Server Hardware


00:00:00   Do you have opinions on kit lenses by chance?

00:00:02   After last show where the three of us were talking about

00:00:07   Apple's TikTok strategy with iOS,

00:00:11   somebody tweeted at us, Pablo Bendersky,

00:00:13   which I probably pronounced wrong, I'm sorry Pablo.

00:00:17   And he wrote a blog post that was actually fairly short.

00:00:19   So hopefully we'll remember to put this in the show notes

00:00:21   and you can take a look, but he made an interesting point

00:00:24   and the premise of it was, you know,

00:00:26   what's interesting is that the,

00:00:28   and I'm gonna actually read this here,

00:00:29   The TIC versions of iOS came with the S versions of the devices, iOS 3 with iPhone 3GS and

00:00:35   iOS 5 with the iPhone 4S.

00:00:37   On the other hand, big hardware releases such as the iPhone 4 and iPhone 5 launched alongside

00:00:43   TOC releases of iOS, iPhone 4 with iOS 4 and iPhone 5 with iOS 6.

00:00:50   So now not reading anymore, the general premise he drives at is, "Hey, even if you consider

00:00:57   the software on a TikTok, and if you consider the hardware on a TikTok, what's interesting

00:01:02   is their offset—it's like a 180-degree offset sine wave—so there's always going

00:01:08   to be a Tik and always going to be a Tok, and that's kind of an interesting place

00:01:11   to be in.

00:01:12   And I thought that was an interesting point, so I just wanted to point that out.

00:01:15   JAYLEE MUSIAL WALSH: Weren't we having trouble last show thinking about which iOS release

00:01:19   was big or small, and then we were trying to list off the major features?

00:01:23   I don't know if I buy that theory because it just doesn't seem to me that there has

00:01:30   been that kind of cadence to iOS.

00:01:33   But it could be that I'm just not paying enough attention.

00:01:35   But the fact that we were trying to think about, "Was that a big release?

00:01:38   What was in that one?"

00:01:39   Either we're just old and have bad memories and can't remember what was in them, but they've

00:01:42   all seemed pretty even to me at this point.

00:01:45   Certainly none of them were like a snow leopard, you know?

00:01:47   Right.

00:01:48   And that's kind of the thing.

00:01:49   I think I'm with you on that.

00:01:50   point to any iOS release and say, "Oh, that one was the one that fixed most of the problems and added

00:01:54   a bunch of new APIs but didn't change much to the user."

00:01:58   Maybe six, but I wouldn't say that's enough to really establish a pattern.

00:02:02   At least OS X has had now two ticks and two talks.

00:02:06   It went leopard, snow leopard, lion, mountain lion. Those four releases

00:02:10   I think have very clearly followed that. But even before that, they didn't really.

00:02:14   Well, Jon, you're the expert, did they? Jonathon

00:02:14   Well, John, you're the expert, did they?

00:02:16   No, not really.

00:02:17   I mean, it was just more-- that's

00:02:18   why I think we're still in that early period of iOS

00:02:21   where it's really simple, not a lot of features,

00:02:25   a bunch of stuff doesn't work.

00:02:27   And then the ones that you have to get out

00:02:28   to shore up the API for public use

00:02:31   and then some major features.

00:02:33   Like, it pretty much falls closely along with the 10.0,

00:02:38   10.1, and it's different because it was SDK.

00:02:41   SDK was 2.0, I think.

00:02:44   And then three, four, and five roughly matching with Panther,

00:02:50   Tiger, and Leopard.

00:02:53   And six should have been at Snow Leopard,

00:02:55   but it wasn't because six brought even more

00:02:57   features and more new stuff.

00:02:58   It was definitely not a Snow Leopard type release.

00:03:03   So I still think we're in the early part of iOS's release

00:03:06   cycle.

00:03:07   Maybe seven will be the first.

00:03:10   Or maybe seven will be the--

00:03:13   Right after 7, maybe 8 will be something that's more tame

00:03:16   or whatever.

00:03:16   But it seems like with all the rumors of 7

00:03:18   that it's not going to be a tame release.

00:03:20   Even if it's just visual stuff, I

00:03:22   think that will still stand out in our mind

00:03:24   as, oh, that's when they rejiggered the UI

00:03:25   and rearranged the apps and all that business.

00:03:29   But we'll see.

00:03:30   But I would say there's definitely no snow leopard.

00:03:33   We haven't had a snow leopard in iOS yet.

00:03:36   Now, and the other interesting thing

00:03:38   that somebody on Twitter pointed out,

00:03:39   and I don't have this @ reply handy,

00:03:40   so I apologize to whoever you are.

00:03:42   But they pointed out that none of us thought about Siri

00:03:45   when we were trying to think of kind of marquee features

00:03:48   on different iOS versions.

00:03:49   And the three of us didn't even think about Siri.

00:03:51   And I don't know what that says,

00:03:53   but it's an interesting observation that obviously

00:03:56   didn't cross my mind when we were talking about the OS.

00:03:58   I thought about Siri.

00:03:59   I just didn't say anything because it's like, well,

00:04:01   you know, Siri.

00:04:02   We've talked about Siri in the past.

00:04:03   Because it was tied to the 4S as a device,

00:04:07   mostly for marketing type reasons,

00:04:09   because it was tied to this device,

00:04:11   and that was the release.

00:04:13   It was more like a feature of that phone and all devices

00:04:16   from that point forward.

00:04:17   It was less of like a, OK, starting with iOS 5,

00:04:19   everybody gets this cool thing.

00:04:21   Yeah, that's something they haven't done on the Mac.

00:04:23   It's not like they roll out Time Machine and say,

00:04:25   and if you want to use Time Machine,

00:04:26   you've got to have a Mac that you bought this year.

00:04:28   And that would not fly as much.

00:04:30   And they pulled it off with the 4S as a differentiator.

00:04:33   But it's why we don't associate it with,

00:04:36   oh, that was the OS that had Siri.

00:04:38   It's more like that was the phone that had Siri.

00:04:41   - Yep, fair point.

00:04:42   And anything else on FU, follow up,

00:04:45   whatever we're calling it?

00:04:46   - I don't think so.

00:04:48   - All right, so one of the things I wanted to talk about

00:04:51   today was Marco, you finished revi,

00:04:54   or well, maybe not finished,

00:04:55   but at least got to a stopping point for now

00:04:57   on your PHP framework.

00:04:59   And I didn't know if you had anything to share about that.

00:05:02   And if not, I'm gonna ask another probing question,

00:05:04   but how did that go?

00:05:05   - I can't wait for the probing, but.

00:05:07   (laughing)

00:05:09   - Title.

00:05:10   Now we have a kind of fake, kind of real showbot in the chat

00:05:14   now.

00:05:15   So with the PHP thing.

00:05:16   So I've had this PHP framework that I've built.

00:05:21   I mean, a long time ago, we built Tumblr with it.

00:05:24   And before Tumblr, we built-- David Karp and I

00:05:27   built some client applications back

00:05:30   when we were consultants before David decided to start Tumblr.

00:05:33   So we had this PHP framework.

00:05:36   And it's a basic MVC framework.

00:05:38   And the reasons we wrote our own back then

00:05:41   were that we had tried Rails, and we had tried Cake for PHP.

00:05:44   And back then-- this was in 2006-- we didn't-- at least

00:05:51   I didn't, and I convinced David--

00:05:53   I didn't like all of the behind the scenes magic

00:05:56   that frameworks did unexpectedly.

00:06:00   Some magic is good, and some magic is helpful.

00:06:04   My theory was more that I wanted the framework

00:06:06   to be more like libraries and less like at the time what

00:06:11   was trendy for, quote, frameworks, which is they

00:06:14   would do so much for you.

00:06:16   And so it was hard to tell.

00:06:19   If you were seeing a certain header being

00:06:21   set or certain behavior or certain filtering

00:06:23   or certain bugs, it was hard to find where the heck that

00:06:26   was coming from in the code.

00:06:27   And this was especially a problem in Rails

00:06:30   because of Ruby's mixins and that becoming especially

00:06:34   trendy in 2006, so everyone was kind of overusing them.

00:06:39   And so it was very, very-- much like categories in Cocoa--

00:06:43   it was very, very hard to find the code you were looking for,

00:06:47   to manage what was going on, because there

00:06:49   was so much magic happening all the time with everything.

00:06:52   So my theory was I wanted something

00:06:56   that could provide helpful features, like a library,

00:06:58   but where I knew everything that was going to happen

00:07:02   and what was not going to happen.

00:07:04   So it was very clear to me, OK, me

00:07:06   as the person writing this action or this model,

00:07:09   what is my responsibility?

00:07:10   What will happen and what won't?

00:07:13   And so we wrote our own very basic framework

00:07:16   with some lessons from Rails, some lessons from ourselves,

00:07:19   some lessons from Cake.

00:07:20   And we basically matured that into Tumblr.

00:07:26   And as Tumblr went on, and it is BSD licensed,

00:07:29   but it's not distributed anywhere.

00:07:32   So our clients from before Tumblr, they all have a copy.

00:07:35   A few programmers here and there have a copy.

00:07:37   I have a copy.

00:07:38   Instapaper is running on a copy.

00:07:40   And the magazine's running on a copy.

00:07:42   It's this not really open-sourced framework,

00:07:45   but it would be open-sourced if it ever actually got out

00:07:47   legally.

00:07:49   So what I did, when I left Tumblr, I forked it.

00:07:52   And so Tumblr continued doing their own thing with it.

00:07:55   And Tumblr did tons of changes to it,

00:07:58   because they had to, because they grew like crazy.

00:08:01   And they had to build in automatic sharding support

00:08:03   and stuff like that.

00:08:04   All sorts of stuff I've never seen.

00:08:06   And with Instapaper, I didn't really

00:08:09   have to do anything to it.

00:08:10   So it mostly just sat there.

00:08:12   And I exchanged a few things here and there,

00:08:15   but I didn't really pull anything out.

00:08:16   So it still had a whole bunch of stuff

00:08:18   from halfway through Tumblr scaling

00:08:20   that I had put in there.

00:08:22   And it was also based on PHP 5.1 and 5.2.

00:08:27   Even though we ran it mostly in 5.3,

00:08:29   the late static binding support in 5.3 that was added, which makes static classes work in a useful way for the first time in PHP.

00:08:36   I don't know why they didn't have that at the beginning, but anyway, that's a little too specialized for this show.

00:08:42   I was able to modernize the framework if I wanted to, but Instapaper's codebase at that point was big enough that it was kind of unwieldy to do that, and it wasn't really ever worth my time to do that.

00:08:54   Having just sold Instapaper though, I wanted to start more web stuff in the future, and

00:09:03   I wanted to have a really solid foundation for doing that.

00:09:06   And I thought about using another language, but as I've discussed in this show before,

00:09:10   much to many people's irritation, I don't think Ruby or Python really provide enough

00:09:18   of a benefit over PHP, that if you're already an expert in PHP that you really need to become

00:09:22   an expert in one of those.

00:09:23   I think you can pick any of those three,

00:09:26   become an expert in one of them, and you're pretty well set.

00:09:29   You don't really need the other ones,

00:09:30   unless you really want to learn a lot more languages

00:09:33   for other reasons, like personal development or market value

00:09:35   or whatever else.

00:09:36   But if you're short on time, and if you don't really

00:09:39   want to be learning tons of languages

00:09:40   and you'd rather take one and master that one,

00:09:42   I don't think you need to master more than one of those three.

00:09:46   And I think the alternatives, things like Node,

00:09:50   I would be looking at those in probably the future,

00:09:52   but I think they're a little bit early right now.

00:09:54   So I'd rather not go that way right now.

00:09:56   So I figure, you know what, PHP is great for me.

00:09:58   I know a lot of people hate it.

00:10:00   But PHP is very good for me.

00:10:02   I know it extremely well.

00:10:04   And I know how to scale it.

00:10:06   I know how to run it very cheaply and very easily

00:10:09   with very little effort and very few 3 AM wake ups

00:10:11   for the sys admin, which is usually me.

00:10:14   And so I decided, you know what, there's nothing really

00:10:18   inherently wrong with PHP that overcomes

00:10:22   the barrier for me to master something else to this degree

00:10:26   when I don't really want to be writing big web

00:10:30   services as the only thing I do.

00:10:32   If that was the only thing I was going to do, that would be fine.

00:10:35   I would be very, very happy with it.

00:10:37   And I would learn lots of different languages

00:10:40   and pick the best one and really dive deeply into that one.

00:10:43   But for me, the web service-- and actually,

00:10:46   there's a great blog post by Brent Simmons

00:10:47   that while I'm rambling on, you guys should read this.

00:10:50   And I'll talk about it in a little bit.

00:10:51   I pasted a link in the chat.

00:10:55   For me, making web services is really a supporting role

00:10:58   to what I really want to do, which

00:11:00   is making whole products that are mostly apps.

00:11:05   For me, Instapaper pretty much was that.

00:11:08   The Instapaper web service was very low needs,

00:11:11   and I devoted very little to it, for good and bad,

00:11:14   because what I really wanted to be doing

00:11:16   was working on the app.

00:11:17   And the app was the premiere experience.

00:11:19   Well, whatever else I do in the future

00:11:21   is probably going to be that same way.

00:11:23   I mentioned in the past I have a few prototypes of things

00:11:26   that I'm trying to figure out what I want to do next with.

00:11:29   And they all have web components.

00:11:31   But all the web components are relatively unimportant

00:11:35   compared to the app.

00:11:37   And so for me, it's not worth learning entire new web

00:11:41   languages and mastering those platforms

00:11:43   and learning the hard way how to scale them all

00:11:47   and spend all that time and stress on that

00:11:49   when I'd rather be putting that effort into the apps.

00:11:52   And just the website, you know, I

00:11:54   don't need the website to be my experimental playground.

00:11:58   So anyway, so I decided this was a great time

00:12:05   to modernize my framework.

00:12:07   Because I had only the magazine using it,

00:12:09   and the magazine's code base is very, very small.

00:12:12   So I thought this is a great time to modernize it.

00:12:14   What I really need, though, is a test app.

00:12:17   And that way I can kind of write the test app

00:12:18   as I modernize it so I know what I'm breaking,

00:12:20   I know how I'm breaking it, and I know what it needs to be.

00:12:23   And I can kind of play with the APIs as I'm making them,

00:12:27   and as I'm changing them to see like, okay, what works,

00:12:29   what doesn't, what ends up being more or less elegant

00:12:32   than I expected, et cetera.

00:12:34   So, and my goal here is also to open source

00:12:38   this framework now, to finally make it

00:12:40   so incredibly different from what Tumblr is running that,

00:12:44   I mean, legally I could have done it already,

00:12:48   but I was always a little afraid to release the framework that

00:12:52   runs Tumblr.

00:12:54   What if somebody finds some security problem

00:12:56   and exploits it against Tumblr?

00:12:58   I would really not feel good about that.

00:13:00   So I was always very afraid about that.

00:13:03   I'm less afraid when it's my own thing.

00:13:05   I at least know that, OK, I'm willing to do that to myself.

00:13:10   I'm less happy about doing that to somebody else,

00:13:12   especially somebody as big as Tumblr.

00:13:15   So I thought, I would love to open source this.

00:13:18   And I would love to modernize it.

00:13:21   And I would love to also write the sponsor tracking system

00:13:24   that I've been meaning to do for markers.org for months.

00:13:28   Because the way I track my sponsors for markers.org

00:13:30   is this giant spreadsheet that sucks.

00:13:32   And it's so human-based and error-prone.

00:13:38   And it always freaks me out.

00:13:39   What if I book the wrong-- what if I publish something

00:13:42   from the wrong day or the wrong week?

00:13:44   What if I think a sponsor hasn't paid, but they have?

00:13:48   or vice versa, like what if I type a number wrong,

00:13:51   if I-- just a typo and something is really messed up.

00:13:55   So I needed something better than a spreadsheet

00:13:58   to manage that.

00:14:00   I also, once we started this show

00:14:02   and we started having sponsorships for this show,

00:14:04   then I had to manage, OK, well, we sell sponsorships for this.

00:14:07   That's another property to sell sponsorships for.

00:14:09   So initially, it started out as just another table

00:14:11   in my spreadsheet that was very similar to the first table.

00:14:14   But then we have to build on things

00:14:16   like, we all split the money that comes in, so I had to build in how to pay-- I take the

00:14:23   money in and I have to pay you guys out your portions of it.

00:14:26   So I have to-- that's another layer on top of this that is error-prone and could be messed

00:14:33   up and starts really pushing the boundaries of what spreadsheets can do gracefully.

00:14:37   I mean, they can do it in many non-graceful ways, but what they can do gracefully is--

00:14:42   this is kind of a bad thing for them.

00:14:46   And then what really killed it is I wanted all three of us

00:14:51   to be able to log in and look at sponsors and sell sponsorships.

00:14:55   And so we needed multi-user access.

00:14:57   And then you can go Google Docs, but then it's just--

00:15:00   the more layers you add to this spreadsheet,

00:15:02   the crappier that solution becomes,

00:15:04   and the more likely it is, like exponentially more likely

00:15:07   it is, that something will go wrong at some point,

00:15:10   and that's really bad.

00:15:12   So I decided, let's build this system.

00:15:17   And that's a perfect test case app for my new framework

00:15:22   or my new version of this framework,

00:15:23   which is almost completely rewritten, it turns out.

00:15:27   And it's way smaller than the old framework.

00:15:30   And my goal is to open source both the framework

00:15:33   and this example project of my invoicing system fairly soon.

00:15:39   So, do you guys have anything to say after that massive tirade?

00:15:44   We're like 23 minutes in now.

00:15:46   Well, I do, but I'll give John a chance, because I have a comparatively much shorter monologue that I'm queuing up.

00:15:52   I'm looking at your notes, so I know what you're going to ask him, and it's exactly the same thing that I'm going to talk to him about.

00:15:57   Oh yeah, the notes. I should open that up.

00:15:59   Go ahead and steal my thunder.

00:16:00   So go for it. Yeah, the notes that apparently Casey and I use, but Marco shunned.

00:16:03   Oh, I got it. Sorry. It was in my Chrome ghetto.

00:16:06   Yeah, that's where mine is.

00:16:07   So since I'm talking, I will go ahead and ask by telling a very brief story.

00:16:11   So a few days back, Marco sends John and I an email saying, "Hey, this is up.

00:16:16   Here's where you can go."

00:16:18   And I load the URL and all I see is a text form and a button.

00:16:24   And the text form says, "Give me your email address and then log in."

00:16:30   And I think to myself, "Well, this is weird.

00:16:31   I haven't created a user account," or knowingly anyway.

00:16:35   I haven't given him a password of any sort.

00:16:39   What's going on?

00:16:40   So I enter my email address, I hit the button, and next thing I know, Marco's fancy sponsorship

00:16:46   tracker says, "Okay, look at your email.

00:16:48   You'll get a login URL there."

00:16:50   And so, sure enough, instantly I look at my email and there's an email from Marco's fancy

00:16:57   app thing.

00:16:58   Then it says, "Here, click this link and you can get logged in."

00:17:00   And that's exactly what happened.

00:17:02   And so it was a very different and interesting way to handle user authentication and authorization—I

00:17:09   always get the A's wrong—but handle logging in without me ever giving him a password.

00:17:14   So in summary, my question to you, Marco and John, feel free to jump in, is why do you

00:17:18   hate passwords so much?

00:17:19   Well, he does the same thing on the magazine site, which I'm assuming was the first place

00:17:24   you did this?

00:17:25   Busted.

00:17:26   Yeah.

00:17:27   I have the—when I saw it, I immediately knew that you were doing the same thing as

00:17:30   the magazine site, or it seemed like it.

00:17:32   Yeah, most of the same code, too.

00:17:33   Yeah, and I didn't like it either.

00:17:35   I complained to my wife that he's doing this thing

00:17:38   with no passwords again.

00:17:40   One of my first interactions I had with you and your products

00:17:44   online was when I first signed up for Instapaper.

00:17:46   And back then, you didn't need a password for Instapaper,

00:17:49   which was awesome, and I love it.

00:17:51   Not like this.

00:17:52   You just literally didn't need a password.

00:17:53   So for the first, whatever it was, year or two years

00:17:56   or however long you let us not have passwords,

00:17:58   anybody could have seen my Instapaper link.

00:18:00   'cause I didn't care, I wasn't putting anything secret there.

00:18:02   I just loved not having another password to remember.

00:18:04   I just had to remember my username,

00:18:07   and there was my Instapaper, and I didn't care,

00:18:09   that was totally insecure, fine, go ahead.

00:18:11   I guess someone could have used that

00:18:12   to start adding stuff to my Instapaper,

00:18:14   and then you had the option to add a password,

00:18:16   so if someone found it and started abusing it

00:18:18   and erased all my Instapaper links,

00:18:20   I'd be like, all right, fine, I'll add a password.

00:18:21   So I'm all thumbs up for not having a password,

00:18:24   but thumbs down for every time I log in,

00:18:27   sending me an email with the link.

00:18:29   Because first of all, when you first did this in the magazine,

00:18:31   there was like mail backlog.

00:18:33   And nothing is more frustrating than wanting

00:18:35   to get into something and not being able to because your mail

00:18:37   hasn't arrived.

00:18:38   Well, yeah, that was a one-time issue.

00:18:40   Right, but even when it's not your fault.

00:18:42   Say it's on the receiving end.

00:18:43   You know, email is stored and forward.

00:18:45   It could be not in your control that I'm sitting here--

00:18:48   Could be green listing.

00:18:49   --bringing on my refresh button and trying

00:18:50   to wait for my thing to come.

00:18:51   And the second thing is, this is really mostly iOS's fault,

00:18:54   but this is the environment you're in.

00:18:56   I use the Gmail app on iOS.

00:18:58   And when the email comes with a link on it,

00:19:01   I can't tap that link because then that will open

00:19:03   in the Gmail app's built-in browser.

00:19:05   And I'll be logged in there in some sort of right-hand side

00:19:08   navigation pane to the Gmail app.

00:19:10   But I'm not logged in in Mobile Safari.

00:19:13   And that link is now dead because I

00:19:14   think you kill them off after they get used.

00:19:16   I don't think an hour had passed, right?

00:19:18   And so now I have to go back to the site,

00:19:20   make an email on me another link,

00:19:22   remember to tap and hold on it and copy the URL,

00:19:25   then go to Mobile Safari, then paste that into a thing,

00:19:27   and then use it and then I'm logged in over there.

00:19:30   So I do not find the user experience

00:19:33   of this passwordless thing agreeable.

00:19:36   - Well, and it's funny because I actually don't have

00:19:40   any problem with the passwordless idea in principle,

00:19:45   but I too was like, wait, I have to wait

00:19:47   for a darn email to come in.

00:19:48   And in retrospect, thinking about it more,

00:19:50   I actually think by and large, this isn't a bad setup.

00:19:54   And it strikes me as a, and I'm not a security minded,

00:19:57   you know, tin foil wearing nut job, but it strikes me as though it's a pretty good

00:20:01   setup.

00:20:02   It's the same as using a password, because if you have access to my email

00:20:05   account, passwords don't matter, and if you have access to my email account for

00:20:08   this thing, passwords don't matter. It's not a security thing where it's any lesser or more

00:20:11   secure than the other thing. Well, it is in one way, and that is that

00:20:15   you can't guess somebody's password because they don't have one. The only way

00:20:18   to log in

00:20:19   is by having access to their email account. Well, you just guess the password for their email

00:20:23   address. I mean, you're just moving, you know,

00:20:25   Moving the password that they have to know from your site to their email provider.

00:20:29   That's true, but that is a little bit better though.

00:20:30   Especially since so many emails have two-factor now.

00:20:33   Exactly.

00:20:34   Took the words right out of my mouth.

00:20:37   If you support a Google login and I can use my existing two-factor to get into your site,

00:20:42   then there you go.

00:20:44   Yeah, well.

00:20:45   Oh, well.

00:20:46   Yeah, no, it doesn't matter for this because it's a sponsorship website.

00:20:49   Obviously here, but I've been thinking about it because everyone always wants to solve

00:20:52   the problem of identity on the web because it's such a pain in the butt no matter what

00:20:56   you do. There's just no good solution. Everything has trade-offs, and the trade-offs for this

00:21:01   one are just a different set of trade-offs. Right.

00:21:04   But I find it particularly—especially for the magazine. I'm subscribed to the magazine.

00:21:08   I get it in the iOS app. Occasionally, I want to read it on the web thing, and inevitably,

00:21:12   I somehow find that I'm using a browser on a machine that doesn't have the right cookie,

00:21:16   and I got to go through the email thing. That tiny delay makes me unhappy. I'd rather

00:21:21   have it like, you know, auto fill my password or just remember what my password is or use

00:21:26   some other means that I can get myself logged in to read the magazine articles because,

00:21:30   you know, it's perpetually telling me sign up for the magazine, subscribe, you should

00:21:33   do like, no, I do, just show me the article.

00:21:36   It keeps you logged in for like a year.

00:21:38   I know, but I don't know, somehow, I mean, it works, I'm always hosing my cookies, right,

00:21:43   because you're doing web development and I have seven browsers and I'm constantly wiping

00:21:46   cookies and doing private browsing and I don't know, maybe I just have to spread, maybe it's

00:21:51   It's like two-factor, where I have to have some sort of ramp-up period where I spread

00:21:54   the little authentication tokens to all of my browsers and all of my cookie jars, as

00:22:00   we called them back in the day.

00:22:03   But it hasn't happened yet.

00:22:04   Okay.

00:22:05   Casey?

00:22:06   To be honest, I don't have a whole lot to add.

00:22:10   I don't mind it nearly as much as John does, but I'm not nearly as much of a curmudgeon,

00:22:15   apparently.

00:22:16   I mean, I think it's interesting because we all hate passwords. We all want them to go away.

00:22:20   But obviously nobody's really cracked that knot on what the right answer is.

00:22:24   And for me, since this is a site that I'm not going to look at every single day,

00:22:28   I don't find it as egregious a pain in the butt.

00:22:32   But with that said, if it was something that I used frequently,

00:22:36   I think I would find it just as annoying as John does.

00:22:40   Yeah, I think it's a very hard problem. The magazine was—I never did this on Instapaper.

00:22:44   paper. The magazine was the first place I did this publicly. And the way it works, in

00:22:49   case someone is still behind on this, it basically treats every login like a password reset,

00:22:56   where you say you want to log in, type in your email address, and it emails you a link

00:22:59   with a single-use hash on it, and then the app trades that hash for a login token and

00:23:05   you're logged in. But it's like a password reset where you don't even set the new password,

00:23:10   it gets logged in. And it's interesting. It solves, very similar to Instapaper's original

00:23:21   no password accounts at all, it solves some problems, but it creates other problems. And

00:23:28   so I'm not sure if it's a net win in general. I think you can look at, you know, where I've

00:23:34   used it here, which is kind of like an admin panel. I think admin panels, any kind of intranet

00:23:39   type thing where security is needed.

00:23:44   Anything like that where it's not the general public having to

00:23:48   log in and you don't have a ton of people using the Gmail app on

00:23:51   their phone and not knowing how to copy links and stuff like that.

00:23:55   Situations like that, I think it's a great solution.

00:23:58   I will probably do this for every admin panel that I ever do again

00:24:04   because it's so much more secure than just having some kind of admin password up there.

00:24:08   And it's just great for that. And I was thinking…

00:24:13   Well, what makes you think it's more secure? Because I can see how it's better from your

00:24:17   perspective as the person running the service, because you don't have to store passwords. And

00:24:20   if you don't have to store passwords, that means the pressure is off of you to keep those passwords

00:24:24   secured and all that business or whatever. But I don't know if that's better for the people who

00:24:30   are using the product because as far as they're concerned, it's your responsibility to make

00:24:34   sure the password, the encrypted versions of the password don't get out, and make sure

00:24:38   you use bcrypt on the slowest setting or whatever the hell crazy stuff you're supposed to do.

00:24:45   I think this definitely makes it better for you writing the service, but as a user of

00:24:48   the service, I don't see any security advantages from my perspective.

00:24:52   I just see the slowdowns.

00:24:53   Well, I think the security advantage is that my service now becomes as secure as your email,

00:25:00   which is probably more, given that most people use webmail for their own email addresses,

00:25:05   like not for work, but most people use webmail and most webmail services are really locking

00:25:10   down their security really well. I don't think it's that bad of a thing to rely on,

00:25:16   because as you said, most things are only as secure as your email at most anyway, because

00:25:22   you can always do a password reset. So if I were to do any kind of set your own password

00:25:29   for this, even if I would add two-factor or something,

00:25:32   if I have some way to use your email address

00:25:36   as a reset mechanism, for the most part,

00:25:39   unless I have my own two-factor thing,

00:25:42   which is possible, but I think that would be

00:25:44   even more overkill.

00:25:45   Like, if you don't like having to wait for an email

00:25:47   and click a link, you're gonna hate two-factor.

00:25:50   - But if you're gonna delegate authentication

00:25:51   to someone else, 'cause like you're delegating

00:25:53   to our email provider, why not like literally

00:25:55   delegate authentication to someone else

00:25:56   and do what those sites do?

00:25:58   Either you pick one that you decide to be blessed, which is fine for an admitted site,

00:26:01   or you have three buttons that say "log in with your Facebook ID", "log in with your

00:26:04   Google ID", "log in with your OpenID".

00:26:05   In all those cases, you're delegating the authentication to an external service, but

00:26:09   then you don't have to deal with it at all, except for writing the implementation of those

00:26:12   things.

00:26:13   And for an admin page, you would say, "Okay, well, you can't log in with Facebook, and

00:26:16   you can't log in with OpenID, but you can log in with Google."

00:26:20   You know, because you can limit your options there.

00:26:22   First of all, this is simpler and easier for everybody.

00:26:25   See, the big problem with those login with Facebook or Twitter or whatever,

00:26:30   with those things, and with OpenID, which is itself a massive train wreck of normal person usability.

00:26:35   Even for geek usability, OpenID is a mess.

00:26:40   Look at Stack Overflow. They were OpenID only for the first few years of their life.

00:26:45   And because Jeff Atwood really believed it.

00:26:50   But the problem was, even with that extremely nerdy audience,

00:26:51   it was still a major problem that caused tons of support.

00:26:54   - I think the environment has gotten a lot better.

00:26:56   Like the Teespring site. - Open ID?

00:26:58   - Doesn't the Teespring site use Open ID?

00:27:00   'Cause when it's-- - It offers it as an option,

00:27:01   but I just created my own account.

00:27:02   See, here's the major problem with those kind of things,

00:27:05   where you can say, log in with Open ID,

00:27:07   or with Facebook, or with Twitter.

00:27:09   The main problem is that every time you go back

00:27:11   to that site, you have to remember which one

00:27:13   you logged in with the first time.

00:27:14   - Well, that's what I was saying, just offer one.

00:27:16   Like for an admin site, you just pick one.

00:27:17   - Well, that's what I do.

00:27:18   It's your email address.

00:27:20   (sighs)

00:27:20   I mean, I don't see how it's different.

00:27:22   - When I go to the Teespring site,

00:27:24   I click the button that says log in with Google,

00:27:26   and I don't have to do anything further,

00:27:27   because I've already done the thing once.

00:27:28   When I go to your site, every time I have to click,

00:27:31   and go get the email, and you're like,

00:27:32   well, it should still have your cookie, but I don't--

00:27:34   - But when you go back to the Teespring site in a year,

00:27:37   and you haven't used it in that entire year,

00:27:39   and you go back here, and it's, oh, what did I log in with?

00:27:42   Which one was it?

00:27:43   You know, it's a problem,

00:27:44   and that creates support emails like crazy.

00:27:47   You would not believe how much support that generates

00:27:49   when you have that kind of option.

00:27:50   But again, for an admin site, support is not an issue.

00:27:52   Well, true.

00:27:53   But I would also argue that for an admin site,

00:27:58   there's really no benefit to doing that your way instead

00:28:02   of my way for an admin site.

00:28:03   Because either way, you're likely to hit a small delay

00:28:08   sometimes either way.

00:28:10   If you're saying log in with my Google account,

00:28:12   yeah, it might be one click.

00:28:13   Or you might have to relog into your Google account

00:28:16   because it might have been too long since you've last

00:28:18   logged in.

00:28:18   Or it might want to reauthenticate you.

00:28:19   The Google thing has some sort of timeout.

00:28:21   If you're making your cookies last for a year or whatever,

00:28:24   that means anyone who finds them-- it doesn't matter.

00:28:27   But if I log in, I have to remember

00:28:29   that if I log in on a strange machine,

00:28:31   I have to make sure that I get rid of that so that that person

00:28:34   can't perpetually log in as me because now they've

00:28:36   got the cookie.

00:28:36   You know what I mean?

00:28:37   Whereas when you're delegating entirely to an external service,

00:28:40   all the management of revoking access and timing out

00:28:43   and whatever is on the external service, right?

00:28:47   Yeah, but also, the extra level of complexity of dealing with someone else's service for

00:28:53   this, it adds almost as many annoyances and flaws and support headaches as doing it the

00:29:00   way I'm doing it with emails.

00:29:01   That's what I'm saying.

00:29:03   So what has it been like for the magazine?

00:29:05   This Mint site is a bad example, because there's three people using it.

00:29:08   Right.

00:29:09   Whatever.

00:29:10   But for the magazine, there's many, many more people using it.

00:29:12   I guess you don't have anything to compare it to, because you haven't implemented,

00:29:15   you know, log in with your Google account as well. But what has that been like for,

00:29:21   because you mentioned the copying and pasting from the Gmail app or whatever. Is that just

00:29:24   because I mentioned it, or do you actually get that from Google?

00:29:26   No, that's the preferred of it, actually. I should probably use Gmail at some point

00:29:29   in my life, but oh well. No, with the magazine, it's been interesting. This is why I'm

00:29:38   not sure I would use it again for like a big consumer-facing thing, at least not yet. Because

00:29:43   The main problem is that when you have a username and password, you can enter those no matter

00:29:51   where you are, no matter what app or context you're in, you can enter those immediately

00:29:56   as you say, you can enter those and not have any kind of delay and just be logged in.

00:29:59   As long as you know the password.

00:30:01   And yeah, there's people like us who use one password and generate long strings of garbage,

00:30:04   so for us it's a little bit more complicated, but most people don't do that, unfortunately.

00:30:08   So the big problem is, first of all,

00:30:12   you're tying something to email delivery, which,

00:30:14   similar to tying something to someone else's account,

00:30:16   there are opportunities for that to go wrong or to be delayed.

00:30:19   And email delivery can be somewhat assumed to occur.

00:30:25   You can generally assume that an email will usually get there.

00:30:28   There are, of course, complications there,

00:30:29   but usually you can say an email will get there.

00:30:31   What you can't guarantee is that an email will get there

00:30:34   quickly, because there's things like gray listing.

00:30:37   and all sorts of weird anti-spam measures and server

00:30:42   configurations and everything that will frequently

00:30:44   delay a message by like a half hour or three hours

00:30:48   from the first time it's hearing from you or something.

00:30:50   So that becomes harder when you say, all right, log in.

00:30:54   Click here, and you'll get an email in two seconds.

00:30:56   Well, that's great if it comes in two seconds every time,

00:30:58   but it doesn't always.

00:30:59   It comes in two seconds for most people,

00:31:01   and then there's like the 1% that it

00:31:02   doesn't come in two seconds for, and they get upset.

00:31:05   And they email your support, and there's really

00:31:07   way for you to do anything about that quickly.

00:31:09   And they keep entering their email address and clicking the button again and again, queuing

00:31:13   up seven emails.

00:31:14   Right.

00:31:15   Which keeps invalidating their previous one.

00:31:16   Right, exactly.

00:31:17   And then they try to click the third one that came, not realizing that it was invalidated

00:31:19   when you sent the eighth one.

00:31:22   Exactly.

00:31:23   So it's not perfect in that regard.

00:31:26   And also, this was actually one of the last Build and Analyze episodes.

00:31:30   I talked about when I made the magazine how in the effort of--

00:31:36   or with the goal of simplicity, I

00:31:38   wanted it to have no settings in the app,

00:31:40   just no settings screen.

00:31:41   Like to have everything just be available in the interface

00:31:44   with no gear icon with a settings screen on it.

00:31:47   I didn't want some big long list of check boxes or whatever.

00:31:49   I just wanted it to be very, very simple.

00:31:51   And the problem was that to avoid having a settings screen,

00:31:54   I had to make other bad design decisions,

00:31:57   like how to log out from Instapaper

00:31:59   if you had logged in to send stuff to it. There were other decisions I had to make that

00:32:03   were bad decisions in order to support this one thing I thought was a good decision.

00:32:09   And so the passwordless login system that uses email has a few of those things. There's

00:32:16   a few weird little stupid things that I had to do to make that work that are bad design

00:32:22   decisions, even though the email thing I think is overall a good one standing on its own.

00:32:28   So one of those things is when you want to connect,

00:32:32   when you want to sync your account between the app

00:32:35   and the website on your iOS device,

00:32:37   or when you want to sync your account

00:32:38   between the app and the website,

00:32:39   regardless of where you want to be browsing it.

00:32:41   Like if you subscribed in the app

00:32:44   and you want to browse it on the website on the desktop,

00:32:48   or if you subscribe on the website

00:32:49   and then you want to read it in the app.

00:32:51   Communicating between the browsers and the website

00:32:54   an Apple purchase thing in these three different places,

00:32:58   when you have a password, it would be easy.

00:33:00   It would just be enter this username and password

00:33:03   in the other place, and you're logged in, done.

00:33:06   The way I do it though, since there's no passwords,

00:33:09   is like for the app to register its subscription

00:33:12   with the website, I have the app compose an email message

00:33:16   with an attachment that encodes a bunch of data.

00:33:19   And then it sends an attachment to my server,

00:33:21   The server associates it with the email that sent it.

00:33:24   Which has its own set of problems, like,

00:33:26   well, what if the email that you sent that from

00:33:28   is not your primary email address?

00:33:29   Like, there's other problems with that.

00:33:30   Yeah, I think I did that a few times.

00:33:32   I was very confused on launch, because I don't use--

00:33:35   the email address I use for signing up for web things

00:33:37   is different than the email address I use as my iCloud

00:33:41   Apple ID.

00:33:42   Exactly.

00:33:43   And I wouldn't say that's a common case,

00:33:46   but there's certainly enough people

00:33:47   who do that that it's problematic

00:33:49   if you don't support that in the way they expect.

00:33:54   And so there's that issue, going that way.

00:33:57   And in the other direction, I would

00:33:58   log into the website on your device

00:34:00   and open this link that I'll mail to you.

00:34:02   And then it'll send you the magazine app colon slash slash

00:34:06   garbage URL to open up in the iOS app

00:34:10   to associate it in the other direction.

00:34:12   And that's kind of ugly and can have some of its own problems

00:34:15   as well.

00:34:16   So I had this technical need to associate these accounts

00:34:22   between these two places, and passwords really

00:34:23   would have made that a lot easier.

00:34:25   Well, as soon as Apple comes out with the web service

00:34:28   APIs for iCloud authentication, you'll be all set.

00:34:30   We'll talk about that after the first sponsor.

00:34:33   But yeah, so I think overall, I think the passwordless system,

00:34:38   it's interesting.

00:34:39   And that's why I did it, because it's interesting.

00:34:41   I do think I'm going to keep it for admin panel type stuff.

00:34:45   I don't know if I'm going to launch anything else with that

00:34:47   as the main front end thing.

00:34:48   Because, in fact, if this was like 2005 still,

00:34:54   and most people were still doing everything on their computers,

00:34:58   and you were logging into a website,

00:35:00   that would actually make way more sense.

00:35:01   And it would be fine.

00:35:02   And I would probably do it, no question.

00:35:04   Now, with everyone doing things on all these different types

00:35:07   of devices, with all these different sandboxed siloed apps

00:35:11   and everything, and everything's running full screen,

00:35:13   there's no good multitasking on iOS,

00:35:15   or no good interactive communication.

00:35:18   There's all these problems that actually now

00:35:20   the system is less practical than it would have been in 2005.

00:35:23   And so I don't think I'm going to do it again

00:35:27   for the main login for something.

00:35:29   And also another problem is that

00:35:32   it doesn't work the way people expect.

00:35:35   And because people are used to username and password

00:35:38   or email and password, they're used to that.

00:35:41   When you put something up there they don't expect,

00:35:43   even if it's simpler to you, in your mind,

00:35:46   even if it's simpler, they're thrown off for a second.

00:35:49   Like, wait a minute, how do I do this?

00:35:51   Where do I put the password?

00:35:53   It slows people down, and people make mistakes,

00:35:55   and they question themselves, and they doubt themselves.

00:35:58   And that ultimately can make all of your goals worse.

00:36:00   Like maybe they don't finish the sign up,

00:36:02   or maybe they don't finish their shopping cart purchase

00:36:04   or whatever.

00:36:05   So it's a hard call if you want to deviate from the norm

00:36:08   something so simple as how do you log into your site

00:36:12   that you should use caution if you're deviating.

00:36:16   On the plus side, I have gotten almost zero support emails

00:36:20   for password related or login related anything. The problem is

00:36:24   if you can't figure it out, you probably will just give up, but

00:36:28   obviously I have no password reset requests, which any support person can tell you

00:36:32   that's probably the most common request no matter how prominent you make that link of

00:36:36   drop password, click here.

00:36:37   What if they type their email address wrong?

00:36:39   There's so many problems.

00:36:41   I don't get any of those for the magazine, which is awesome.

00:36:44   So it does save support by a dramatic amount,

00:36:48   but there are other problems with it

00:36:50   that will give me pause for using it in certain contexts.

00:36:56   Anyway, this episode, our first sponsor today,

00:37:00   is Mac Mini Vault. Go to macminivault.com/try.

00:37:05   This is pretty cool.

00:37:06   Mac Mini Vault, they're basically a Mac Mini

00:37:09   co-locator.

00:37:11   And they've been co-locating Mac Minis in their data center

00:37:13   since 2010.

00:37:15   And then the guys there have been in business since '95.

00:37:18   That's 1900, '95, running ISPs and hosting services.

00:37:23   So they have a giant cabinet, and they

00:37:25   wire it to house a whole bunch of Mac Minis.

00:37:28   And all the level of efficiency and volume,

00:37:30   they're able to place your Mac Mini in a data center

00:37:33   for much less than a standard co-location would cost.

00:37:35   Their plans start at just 30 bucks a month.

00:37:38   And you can use the Mac Mini for pretty much anything

00:37:40   you can think of.

00:37:41   I mean, you know, it's a Mac.

00:37:42   You have control over it.

00:37:42   You can do whatever you want with it.

00:37:45   You can use it as part of your development workflow

00:37:47   for development and testing.

00:37:48   You can even use it for web hosting, file storage.

00:37:51   And if you want, you don't even have to use Mac OS 10 on it.

00:37:54   They'll even load Windows Server on it or Linux

00:37:56   if you request that.

00:37:59   And of course, you can access this Mac Mini

00:38:00   from anywhere on the internet

00:38:02   and you see all the flexibility

00:38:03   that your own server provides.

00:38:05   You get a one gigabit per second network connection.

00:38:08   The data center is redundant and secure,

00:38:10   and there's helpful engineers on staff,

00:38:12   as well as free hands-on hours.

00:38:15   And there's no charge to diagnose and redeploy machines

00:38:18   during the free hands-on hours, which is pretty cool.

00:38:21   So you can send in your own Mac Mini,

00:38:22   or you can buy one directly from them.

00:38:24   And they even have rental options.

00:38:26   You can rent a Mac Mini.

00:38:27   So it's a lot like renting a dedicated server somewhere.

00:38:29   Low cost entry there, it's fantastic.

00:38:32   So anyway, and they also have-- I was really impressed

00:38:34   I was browsing their stuff.

00:38:35   They have-- you can add extra third party RAM or like a Samsung

00:38:40   SSD that's way bigger than Apple's and way cheaper.

00:38:42   So they have-- and they'll install that for you

00:38:44   if you buy the Mac Mini through them.

00:38:46   That's pretty cool.

00:38:47   You can actually deck out a pretty nice machine

00:38:48   there for less than what Apple would charge for the extras.

00:38:51   Anyway, visit macminivault.com/try.

00:38:57   You can start a free demo today, or you

00:38:58   can sign up for an account.

00:39:00   Use the promo code ATP50, that's ATP50 for Accent on a Podcast, 50% off.

00:39:06   It takes 50% off the colocation fees for the first three months.

00:39:11   Great customers there, including the Mediao, the people who make Instacast,

00:39:15   they host their backend thing there, and Crashlytics, the big, powerful,

00:39:20   and lightest weight crashing report solution, which is a previous sponsor of my site,

00:39:25   Mac Mini Vault as part of its testing and development

00:39:28   environments.

00:39:28   That's pretty cool.

00:39:31   They have like 700 Mac Minis now,

00:39:32   over 900 pre-wired slots for the future.

00:39:34   They are well set to expand greatly with all of you

00:39:38   going and creating accounts there.

00:39:39   So go to macminivault.com/try, start a free demo,

00:39:44   and use our promo code ATP50 to get 50% off.

00:39:47   Thank you to Mac Mini Vault for sponsoring.

00:39:50   50%.

00:39:51   That is the new winner for largest percentage

00:39:53   off of a sponsor of the show, I think.

00:39:55   50% for the first three months.

00:39:57   I was about to comment as well that good for them,

00:40:00   for really showing they mean business.

00:40:03   I wonder if the existence of the Mac Mini

00:40:05   contributed to Apple's decision to discontinue the X serve.

00:40:09   Oh, I definitely did.

00:40:10   Like we have accidental tech podcasts,

00:40:12   they have accidental server hardware.

00:40:15   Because I don't think they made the Mac Mini thinking,

00:40:17   this is going to go right in the data center.

00:40:19   But as soon as they made a machine that shape,

00:40:21   or like stick on a hundred of those in 2U rack.

00:40:25   - I kind of feel like the Mac Mini,

00:40:26   you know how like when Steve Jobs introduced

00:40:29   the Motorola Rocker phone, that terrible iTunes phone

00:40:32   that came out before the iPhone?

00:40:34   And you could tell when Steve introduced that,

00:40:36   and Gruber wrote about this a while ago,

00:40:38   you could tell like he just really did not like it.

00:40:40   He did not like the product,

00:40:42   he did not like that it had to exist.

00:40:44   Like the market was demanding something like that

00:40:47   and so they made it kind of reluctantly

00:40:49   And he just kind of like shat it out there

00:40:51   and was OK with it.

00:40:53   I think the Mac Mini was kind of like that for him.

00:40:57   I don't think he gave a crap about the Mac Mini.

00:41:00   And obviously, it's a very low priority for Apple

00:41:04   most of the time.

00:41:05   They hardly ever update the thing,

00:41:06   even when the other things get updated with the same CPUs.

00:41:09   It's obviously a low priority, but they keep it around

00:41:12   because they probably do sell a good number of them.

00:41:14   But it's always kind of seemed like this product

00:41:17   that they reluctantly keep in their lineup.

00:41:19   because people do love it way more than Apple seems to.

00:41:23   - Well, Jobs was probably sore about the cube thing still.

00:41:25   And he's like, "You know, I made this machine

00:41:28   "when it was much better looking

00:41:29   "and you guys didn't want it, but whatever."

00:41:30   Here you go. (laughing)

00:41:32   It's aluminum now, it's small,

00:41:33   it doesn't have the cool glass.

00:41:35   Still 800 bucks, here you go.

00:41:37   - Enjoy it, big whiners.

00:41:39   - Exactly. - But it was much cheaper.

00:41:41   It's true, it's one of those products

00:41:43   he didn't seem to have super duper enthusiasm about.

00:41:45   But other, I mean, that's the weird thing.

00:41:47   Like, if the company is not enthusiastic about a product,

00:41:51   but a certain subset of the customers are,

00:41:53   and it's very clear that a certain subset of customers

00:41:56   is really enthusiastic about the Mac mini.

00:41:58   People who want to stick them in data centers,

00:41:59   and even I've seen people who just,

00:42:01   you know, Mac users who have lots of Macs,

00:42:05   if you just need another Mac to just be an extra machine

00:42:08   for back when Xcode,

00:42:10   does Xcode still support distributed compile?

00:42:11   Margo would know.

00:42:12   - I haven't tried it with the new versions, I don't know.

00:42:15   - But like, you know--

00:42:15   - It was always slower for me.

00:42:17   having a mac mini in the corner to be a media machine and just

00:42:21   you know i just need another mac for this other room so i'll just get a mini or

00:42:24   using it as a server to you know attach your storage to and stuff like that

00:42:28   so i see a lot of enthusiasm for that little machine even though apple like you said does not

00:42:32   seem to be that into it

00:42:34   yeah but i mean but you know everyone always thinks myself included you know we've

00:42:36   thought

00:42:37   a number of times in the past we thought oh this is going to be the year they

00:42:40   kill the mac mini

00:42:41   and it never gets killed it's still it's still sitting there even when it goes

00:42:45   like

00:42:45   What's the longest it's gone without an update?

00:42:47   Like two years, maybe?

00:42:48   Without--

00:42:49   And it doesn't go down in price either.

00:42:50   No.

00:42:50   Like when the new one comes out, it's like, oh, still

00:42:52   really expensive.

00:42:53   In fact, usually it goes up by $100.

00:42:56   Yeah.

00:42:57   Yeah, but it's funny.

00:42:58   I have one for a while too.

00:42:59   Because as you say, people use them as servers all the time.

00:43:02   There's so many little things you could do with them.

00:43:05   So anyway, thanks to them for sponsoring.

00:43:08   Macminivault.com.

00:43:08   So I mentioned before the break, I

00:43:13   wanted to talk about this article by Brent Simmons.

00:43:15   Did you guys get a chance to read it while I was rambling about various stuff before?

00:43:18   I read it during the day, Marco.

00:43:21   Of course you did.

00:43:22   Yeah, I at least scanned it earlier.

00:43:23   Okay, so it's called "30 Minutes to Sync on Inessential.com," Brent Simmons' web blog,

00:43:28   and we will—I said "web blog"—and we will post it in the show notes.

00:43:32   I never say "web blog."

00:43:34   And—

00:43:35   Someone just made fun of Glenn for writing that, so you probably picked that up in the

00:43:38   Twitter stream.

00:43:39   Yeah, but Glenn also says "Capital Web Space Site," and I don't do that.

00:43:44   Glenn says a lot of things.

00:43:45   [LAUGHTER]

00:43:48   Oh, man.

00:43:49   Anyway, so at least vlog didn't catch on.

00:43:54   We can at least be happy about that.

00:43:58   I should also-- you know what?

00:43:59   Before we talk about-- now that we're

00:44:00   talking about stupid words for our new media crap

00:44:02   that we invent as nerds, I've heard a lot of discussion

00:44:06   recently about the word podcast.

00:44:08   I heard there's this great podcast called--

00:44:12   I think it's called Next Market.

00:44:13   I'll have to look this up.

00:44:15   And this guy who does, he's doing interviews

00:44:17   about people in the podcasting world,

00:44:19   talking about the business of podcasting.

00:44:20   And Leo LaPorte famously does not like the word podcast

00:44:25   and uses the word netcast to describe what he does.

00:44:27   Dan Benjamin has said he doesn't like it,

00:44:29   but he still uses it because,

00:44:31   although he tries to do like internet broadcasts and stuff.

00:44:34   There's been a lot of,

00:44:35   a lot of people who are big in the podcast world

00:44:38   have tried to use a different word.

00:44:41   And I think-- what do you guys think of the word podcast?

00:44:46   I think it's fairly terrible.

00:44:48   And I think Apple was super duper lucky

00:44:50   that it caught on the way it did,

00:44:51   because Apple didn't coin it.

00:44:55   But it was so tied to Apple's product, the iPod,

00:44:58   that it was like a happy accident.

00:45:00   I guess we own this space.

00:45:01   The name of our product is practically in it.

00:45:04   I've never been bothered by it.

00:45:06   Not to say you guys are wrong, but it's never, ever, ever

00:45:09   bothered me in any capacity.

00:45:10   And I'm not the kind of person that says, you know,

00:45:14   Tumble log over Tumblr.

00:45:15   Or I might say photocopy.

00:45:17   I might say Xerox.

00:45:18   I might say tissue.

00:45:19   I might say Kleenex.

00:45:21   I never get pedantic about any of those sorts of things.

00:45:24   See, podcast is a lame word, though.

00:45:26   Like, I can see why people who broadcast--

00:45:28   like, because it pigeon holes you into, like,

00:45:30   you make things that people listen

00:45:32   to on iPods or similar devices.

00:45:34   And that is a narrow definition of what could possibly

00:45:37   be the future of broadcast, the future of audio programming, you know what I mean?

00:45:42   Like, that doesn't seem like that should be—

00:45:43   Well, but for a while, "blog" was that kind of word. You know, for a while, saying

00:45:48   that, like, if somebody else told you that you wrote for a blog, that was an insult.

00:45:54   But, like, "blog" has a useful definition. Podcast doesn't. Like, blog, if you look

00:46:00   up on something that tries to find a blog, it's not going to ever call the New York

00:46:04   Times a blog because it's not like a personal single voice publication, time-separated,

00:46:10   you know, linear stream of time-separated posts not divvied up into sections through

00:46:14   multiple authors or whatever. Like, a blog is distinct from, you know, just a website

00:46:19   or writing on the web, whereas podcasts is all-encompassing. If you make an audio program

00:46:23   that's not broadcast over analog radio but distributed through the internet instead,

00:46:27   it's like, well, you have a podcast. And I think that's why people don't like it.

00:46:32   See, I think podcasters don't like it.

00:46:36   I know people love podcasts about podcasts.

00:46:38   Podcasters don't like it because of the pretty bad

00:46:43   connotation it has of being low quality.

00:46:46   'Cause everyone's heard bad podcasts before

00:46:48   recorded by some people talking into their

00:46:51   built-in laptop mics for three hours about nothing.

00:46:55   Well, I guess that's not that far from what we're doing.

00:46:57   (laughing)

00:46:59   - I'm glad you called yourself out on that.

00:47:01   You think that's why—I mean, we should get someone on here who objects to it, who's

00:47:05   a broadcaster by trade, like Leo or whatever.

00:47:08   But it's not so much associated with—

00:47:09   I mean, kind of are we now?

00:47:11   Well, but is it because it's low quality?

00:47:14   Because radio has terrible low quality stuff, like the guy who's on at 2 a.m. or the college

00:47:18   kids who are broadcasting their student station.

00:47:21   Radio is filled with terrible programming.

00:47:24   If you don't believe me, like I said, just wake up in the middle of the night and turn

00:47:26   on the radio.

00:47:27   something being broadcast on many frequencies, and it will not be high-quality content. But

00:47:33   radio doesn't get slammed for that. Like, "Oh, that's okay." It's an amateur who—and

00:47:37   granted, because the airwaves belong to all of us and are controlled by the government,

00:47:41   and they divvy out blah, blah, blah, but you still have student radio stations. Whereas

00:47:45   on the internet, it's even more wide open, and maybe the total volume is higher. Therefore,

00:47:50   the percentage of crap is higher as well. But I really think it's because it seems

00:47:54   like it's as if every single website would have to be called a blog and the

00:47:58   New York Times didn't want to be on the web because they knew they were not a

00:48:01   blog but by putting something up it come to the New York Times blog and it was

00:48:04   like no it's it's you know it's distributing written word online we're

00:48:10   not a blog like and even though podcast doesn't have the same connotations it

00:48:13   does it does lessen it like you wouldn't think that if you if you'd fancy

00:48:19   yourself as the audio equivalent of the New York Times you don't want someone

00:48:22   calling you a podcast. That's my impression. But see—go ahead, Casey.

00:48:27   I just, I feel like somebody linked on Twitter today to some—I think it was a blog post,

00:48:35   actually—where somebody said that nerds ruin everything. And I didn't even read

00:48:39   it because I could tell, or I feel like I could tell, the gist of it within the first

00:48:42   couple sentences, which was basically like, people who are really passionate about things

00:48:46   end up over-analyzing and ruining them entirely. And I kind of feel like that's what we're

00:48:51   I don't know what you mean.

00:48:52   - When it comes to the word, yeah, exactly.

00:48:54   When it comes to the word podcast,

00:48:55   I just, I don't see why it's that offensive.

00:48:58   And I'm not, again, I'm not saying I'm right.

00:49:00   I'm not saying you're wrong.

00:49:01   I just, I don't see why it's so bothersome.

00:49:02   It's just a word to describe something.

00:49:04   And candidly, I have never really associated that term

00:49:08   specifically with iPods.

00:49:09   And maybe that makes me odd or different or weird,

00:49:11   but I don't associate it specifically with iPods.

00:49:14   Just like one of you guys said,

00:49:15   I just associate it with a internet-based radio broadcast.

00:49:19   - Well, that's what happens to any word.

00:49:20   use any word and it just becomes the thing it is, right?

00:49:23   But these people who have objections,

00:49:25   their objections were at the early stages.

00:49:27   I don't know if they're now,

00:49:28   I mean I guess they're still doing it now.

00:49:29   - They are.

00:49:30   - They're still going with Netcast now

00:49:31   because that's just what they do and that becomes,

00:49:32   it's almost like the Netcast slogan from Leo

00:49:36   has almost become like a signature brand

00:49:38   of his network specifically,

00:49:39   'cause as far as I know he's the only one using it, right?

00:49:41   - Yeah, I think if anybody else said we're doing a Netcast,

00:49:43   they would think, like the fans would think

00:49:45   he was ripping off Leo.

00:49:46   - Right, exactly, so it didn't work out,

00:49:48   But the genesis of that was in the beginning.

00:49:50   They're like, oh, I don't want to be thought of as just

00:49:53   this thing that people put on iPods,

00:49:55   because the things that coined that term are not what I'm

00:49:59   doing.

00:49:59   I want to be a broadcaster, and I want to have it.

00:50:02   And then at a certain point, it just continues.

00:50:04   But if you teleported Leo Lepore from the world

00:50:09   before a podcast to now, and he wanted to start one,

00:50:12   I think he would be less reticent to call

00:50:15   what he does a podcast.

00:50:16   I don't know, maybe.

00:50:18   Maybe he would just be stubborn.

00:50:19   I'm kind of with Casey on this.

00:50:21   I think the word is a stupid sounding word, no question.

00:50:25   But podcasts have two main problems.

00:50:29   And one problem, as Dan said in his interview on something

00:50:34   else-- I got a link to it.

00:50:36   I forgot the name of it.

00:50:37   Dan said earlier this week how one of the biggest problems

00:50:42   with podcasts is the barriers involved in getting to them.

00:50:47   Like, if someone has never listened to a podcast before

00:50:50   and wants to-- and hears about them, wants to get into them,

00:50:53   then they have to like-- first of all,

00:50:55   someone has to tell them about them,

00:50:57   because no one knows about them.

00:50:58   Like, in the rest of the world, nobody

00:51:00   knows what the crap a podcast is.

00:51:03   Or at least, for the most part, it's

00:51:05   pretty much out of the mainstream vocabulary.

00:51:09   And so you have to find an app-- you

00:51:12   have to have a smartphone, for the most part,

00:51:14   or listen on your computer, which sucks.

00:51:16   if you have a smartphone, you have to find an app to download or to play podcasts. You

00:51:20   have to find the podcast within that app, and then you have to make time to play it.

00:51:24   And then you have to hear it somehow. You have to put that audio somewhere. Maybe you're

00:51:27   listening to headphones, but if you drive a lot and you want to listen in your car,

00:51:30   then you need a Bluetooth connection or an aux cable or something like that. Technically,

00:51:37   accessing podcasts has a lot of barriers to entry right now. That's one problem.

00:51:43   is relative, if we're going to the second problem, the reason podcasts are as popular

00:51:47   as they are now is because the iPod lowered the barrier of the entry below some critical

00:51:52   threshold beyond which it became possible to be a thing.

00:51:55   That's true.

00:51:56   You know, because before that, it was like, "Forget it. Nobody's going to pull this off."

00:52:00   And the iPod, in general, made it possible for people to take audio from the internet

00:52:05   and put it in a context where they actually want to listen to it. So that was like the

00:52:09   first hurdle. And Dan's dissatisfaction with the current one is like, "Oh, that's great

00:52:13   and all, it made podcasts possible to be a thing that normal people can enjoy. But there's

00:52:17   this next barrier of, if you're interested in a television show, you're much more likely

00:52:22   to say, "Hey, this cool new show I'm watching," and tell your relative the name of that show

00:52:26   with an expectation that they will successfully—and maybe a network—they will successfully be

00:52:31   able to watch that show with no further help from you. Whereas if you tell them the name

00:52:35   of a podcast and they never listen to podcasts, they're not going to make it.

00:52:40   Yeah, that's true.

00:52:41   Yeah, that's true.

00:52:42   Alright, so we are making progress there, but there's still the next leap.

00:52:46   The next leap we can see what it is out there.

00:52:50   It's great that it's possible for people to do it on X scale, but we want to go like

00:52:54   ten hundred times easier still.

00:52:55   That's true.

00:52:56   So, and to get there, that gets me to my second problem.

00:52:59   The second big problem with podcasts is this connotation that they're low quality.

00:53:03   But I think that's something that bloggers had to wait out, the internet had to wait

00:53:09   out.

00:53:10   every media has when it's new, that all the mainstream and the established media of the

00:53:16   previous types all assume or say or outright defame the new thing as being only for amateurs

00:53:25   or inferior. Especially because a lot of times at the beginning they are inferior, until

00:53:32   people learn how to do them well. So I think the word podcast is fine. If we could pick

00:53:39   a new one, that'd be great, but we can't. It's too late." And I think people like Leo

00:53:44   who say netcast or who make their own alternatives, I think they're doing a disservice to the

00:53:50   podcasting world by fragmenting this term. And even Dan sometimes will say "internet

00:53:57   broadcast" or something like that, or even "internet radio" or "internet radio shows."

00:54:01   And I think the problem is that streaming radio exists, and that's something else. So

00:54:07   streaming radio like Shoutcast stations and stuff like that, or even radio-like services

00:54:11   like Pandora, Spotify, and RDO, I think those distort the meaning. If you tell somebody,

00:54:17   "Oh, you should really get into internet radio shows," they might think of those things instead

00:54:23   of podcast. Podcast has a very specific meaning that is a very large category that is very

00:54:29   distinct from those other things and can't really easily be confused with them in the

00:54:34   vocabulary. So I think we have this word, I think we're stuck with this word, and I

00:54:39   think it's not that, you know, people would call them amateurish no matter what you call

00:54:44   them, because they're new, still. And I think it's only a matter of time, just like what

00:54:50   happened to blogs, it's only a matter of time before that just becomes a normal word for

00:54:55   people and they know what you're talking about and it doesn't always have that negative connotation.

00:55:00   And it's happening more slowly with podcasts because the podcasting medium as a whole is

00:55:06   not growing as explosively as blogs did. But it is growing more slowly. It still is growing.

00:55:11   It still is seeping into the world and seeping into the mainstream. It's not exploding

00:55:18   in the course of two years. But I think we just have to wait it out.

00:55:23   I think the internet broadcasting thing and those other terms, that still beats trying

00:55:27   to coin your own new word. Even though it may be confusing, trying to come up with a

00:55:34   generic term for what you do that doesn't pigeonhole you under podcast still beats trying

00:55:38   to come up with a thing like netcast or internet blast or making up some other word. That's

00:55:46   why netcast seems like branding for Twit because it's like, "Oh, well, this is a new word

00:55:49   that's made up." As for the popularity thing, the celebrities have arrived on podcast.

00:55:56   have podcasts now. So that phase of the popularization has begun, at least. The professionals are

00:56:02   here. But the thing is, in general, it's weird that they're not the radio professionals.

00:56:08   You know what I mean? It's not like Howard Stern abandoned his radio show to have a podcast.

00:56:12   He just went to satellite radio, right? And he's just going to retire.

00:56:14   Well, that might be up for debate. We'll see.

00:56:18   But the people who showed up were stand-up comedians, television stars, people who are

00:56:22   are not trained as audio broadcasters, but they're famous people at least, so then you

00:56:27   can get a podcast where a famous person interviews another famous person and it's got name recognition

00:56:31   and stuff. But the thing I worry about with podcasts is the way we all listen to them

00:56:35   hinges on these few, like, the mechanism by which they come to be. We get them onto our

00:56:43   iPods or our phones, and that relies on software created by our phone maker. How much of the

00:56:49   ecosystem right now is entirely dependent on the fact that podcasts are in iTunes. Like,

00:56:54   say the next version of iTunes comes out and podcasts are no longer there, period. They're

00:56:58   gone. Then there would be this mad dash, kind of like the Google Reader thing, to reestablish

00:57:02   that infrastructure somewhere. And we just don't have that kind of like—

00:57:05   Well, they're already gone from the app on iOS.

00:57:08   Well, you know what I mean. Like, you wouldn't be able to go into someplace like—on your

00:57:13   phone or on your Mac or your Windows PC or on basically any device, you can go somewhere

00:57:18   and search based on the name of a podcast.

00:57:20   If you can find the right search box, you can type it in.

00:57:23   But if those facilities went away,

00:57:24   like if they weren't in the iTunes store

00:57:25   and there was no iTunes directory for them,

00:57:28   you had to find like ATP.fm to find this podcast

00:57:31   and get something that can read the feed,

00:57:33   and you had to use a third party feed reader and stuff.

00:57:35   Having that centralized clearinghouse,

00:57:37   we rely on that so much because this infrastructure

00:57:39   hasn't been built out.

00:57:40   And like you said, all right, fine.

00:57:42   Well, then what about when you're going to go in your car?

00:57:43   Oh, now you have another seven infrastructure barriers

00:57:46   to overcome of how do I get this audio that I've got on my iPod or on my phone, but now

00:57:50   I want to listen to it in my car, but my car just has a radio. How do I overcome that?

00:57:55   It's a series of annoying barriers that are preventing it. What you really want to just

00:57:59   be like, someone mentions the name of a podcast. You know exactly where to type that name.

00:58:04   You type it in, and then anywhere you are where you could potentially listen to audio,

00:58:07   you can listen to that thing. We are far from that.

00:58:10   Yeah, that's true. I will give you that. But I think the key to all this is the smartphone.

00:58:16   I really do. I mean, I devoted a lot of the last Bone and Analyze episode talking about

00:58:19   this topic in particular. I think the smartphone is the default device to listen to things

00:58:27   on. It used to be that you'd have the iPod in your pocket and then maybe at work you'd

00:58:31   use a computer or at home maybe use some kind of stereo that uses a disk or something crazy

00:58:36   like that. But I think it's very clear now that the smartphone is the new media player

00:58:43   and not just for portable anymore. And I think we're going to keep going in that direction,

00:58:50   where the smartphone is going to continue to be the center of everything, including,

00:58:54   by the way, including the way that your car not only integrates with your media, but the

00:58:59   way your car connects to the internet. I really do think that's going to go through smartphone

00:59:04   data plans. I think the phone is going to be your terminal and your modem and your media

00:59:09   player to the world and everything will, just like Steve Jobs' digital hub thing back 15

00:59:16   years ago, whenever that was, but now it's your phone and not your computer.

00:59:20   I tend to think you're right. And something you said a little earlier, or implied a little

00:59:25   earlier, made me think, and I'm going to butcher what you said and put words in your mouth

00:59:29   and I'm not going to stop talking so you can't argue with me.

00:59:31   Go for it.

00:59:32   You had said or implied that podcasting isn't really new anymore. And obviously,

00:59:39   this is a much more nuanced conversation than I'm giving the chance to respond to. But I found that

00:59:44   interesting because I've actually had friends of mine, after they caught wind of the fact that

00:59:48   I'm podcasting now, they've said to me, "Hey, what does it take to do that?" Because I've had an idea

00:59:55   for a podcast and I want to do that. What did you do? And of course, I've said, "Well, Marco did all

01:00:00   all the work and I just sit here show up and act like a prima donna. But all kidding aside,

01:00:05   I've had people come out of the woodwork and say, "Hey, I'd really love to do my

01:00:07   own podcast." And what I'm driving at is, it's not new in the sense that this thing

01:00:12   as a medium has been going on for a long time. But I think it's kind of new in the sense

01:00:18   that regular people are starting to grasp this and are starting to wonder, "Hey, can

01:00:23   I get a piece of that pie?" And I don't mean financially, just attention-wise, "Can

01:00:26   And I get a piece of that pie.

01:00:28   And so I almost wonder if we're about to see, maybe not a renaissance, but a newfound

01:00:33   interest in podcasting, where in the same way everyone ended up with a blog, and now

01:00:39   everyone has a Twitter account—I don't literally mean everyone, but I wonder if a

01:00:44   lot of what we would call regular people will end up having podcasts.

01:00:49   And so we'll all be talking into space and nobody will be listening.

01:00:51   It's kind of connected with blog infrastructure because like Squarespace, right? So if someone

01:00:56   asked that, if someone asked me, "Hey, I want to have a podcast," what I would tell them

01:00:59   at this point based only on my secondhand knowledge of Marco doing it is like, it seems

01:01:03   like if you just get a Squarespace site, it's pretty easy to get a podcast up and running,

01:01:08   right? And that's like the same thing with back when we were blogging. It's like, "How

01:01:11   do I get a blog up?" You don't want to tell someone, "Okay, step one, get a shared hosting

01:01:15   plan. Step two, install movable type," right? You wanted to tell them to go to editthispage.com

01:01:20   or whatever the old frontier thing, like one of the first sites that you could just go

01:01:24   to the web page and you just edit it, right?

01:01:26   And of course, the way it is now, it was a gigantic explosion and then reconsolidation

01:01:30   of all these web applications that you could tell anyone to use.

01:01:33   Go to Tumblr, type in the host name you want and a username you want and a password and

01:01:37   click a button.

01:01:38   Great, you're blogging.

01:01:39   Good job.

01:01:40   Like, the equivalent of that for podcasting is more or less here, it seems like, with

01:01:45   the Squarespace type thing, but we haven't seen the proliferation.

01:01:47   I mean, Marco, you set it up.

01:01:49   Was it basically that easy?

01:01:50   just go there and say you want to have a podcast?

01:01:53   It really is that easy. Obviously, Squarespace has been a sponsor of the show for a long

01:01:58   time and has future sponsorships so that we should disclose that. But, yeah, of course,

01:02:03   it's extremely easy there. And they aren't the only people to make it easy. I don't think

01:02:08   WordPress.com, which one is the one that hosts all the sites? WordPress.com? I don't think

01:02:14   they offer this, but there's been a million podcast plugins for WordPress over the years.

01:02:21   You can host a podcast pretty easily in lots of ways.

01:02:23   Right. Are we at the Tumblr phase, though, of podcasting, where if some random person

01:02:28   wanted to have a blog, the reason I would direct them to Tumblr is I would say it's

01:02:33   the easiest. That's basically as simple as it could possibly get. You type three things,

01:02:38   and you are now a blogger. Congratulations.

01:02:40   Actually, I built this Easter egg a long time ago.

01:02:44   Tumblr also supports podcast hosting

01:02:45   if you host the files elsewhere.

01:02:47   I believe if you go to any Tumblr site/podcast,

01:02:52   or maybe /podcast/rss, it will give you

01:02:56   a iTunes-compatible podcast feed of any audio posts that

01:03:01   are externally hosted.

01:03:03   So try that.

01:03:04   Anyway, I built that a long time ago.

01:03:07   I never documented it, I don't think.

01:03:09   but it's probably still there.

01:03:11   But I think, you know, podcasting inherently,

01:03:16   it's like how it's really easy for people

01:03:19   to take good photos that look kind of good,

01:03:21   but it's a lot harder to take good video.

01:03:24   I think podcasting is kind of similar

01:03:26   in that making an audio production

01:03:31   is inherently more work than blogging,

01:03:34   and it takes, like to do it well,

01:03:36   you need a certain level of equipment.

01:03:38   you at least need a good microphone. There's a little bit more barrier just in realities

01:03:44   of making this kind of media that I don't think we're ever going to overcome some of

01:03:48   those barriers. Certainly we're not going to overcome the problem of it being complex

01:03:54   and unusual to be able to make a good podcast that is both good sounding and with compelling

01:04:00   content. I mean, that's always going to be, every media is going to have that challenge

01:04:04   of quality and interestingness and relevance. But I don't think we're ever going to

01:04:10   reach the point of Tumblr where you type in three things and you have a podcast. I don't

01:04:14   think we're ever going to reach that.

01:04:16   JEAN-MICHELLE DOUBLEDO I think we can get pretty close because right

01:04:19   now the main barriers for someone doing that would be you're not going to have a good

01:04:22   microphone because computers don't come with them. But maybe the microphones in your

01:04:26   phone start getting better. Is voice recorder quality on your phone good enough? If we saw

01:04:33   and crazy headsets like Google 4Cs.

01:04:36   Oh yeah, for most people, yeah.

01:04:38   Maybe that would be sufficient.

01:04:39   So the auto quality thing might start taking care of itself.

01:04:42   Yeah, the content problem obviously is never going to end.

01:04:44   But that's not-- I mean, the content problem

01:04:46   isn't solved by better blogging tools either.

01:04:47   People just write what they feel like writing.

01:04:49   It's just like Casey was saying, for the people who he knows,

01:04:53   they're not looking to do anything earth shattering.

01:04:55   But they just feel like, I could do that,

01:04:57   and my 10 friends could listen to it.

01:04:59   I could write a blog.

01:05:00   I could have a Tumblr blog, and my 10 friends

01:05:03   and my mom can look at it and that's enough for me.

01:05:05   And the podcast equivalent to that.

01:05:07   And the barrier to that is like,

01:05:08   your friends and your mom are not gonna go through

01:05:10   the effort to figure out how to get your stupid podcast

01:05:12   onto their phone or iPod because that's just too much.

01:05:15   They'll visit your Tumblr page

01:05:16   and look at the amusing thing you put in there

01:05:18   or just look at your last thing

01:05:19   that you posted to Facebook or whatever.

01:05:21   But podcasts are still a little bit too much of a barrier.

01:05:23   They're like, do I really wanna listen

01:05:25   to this guy's podcast?

01:05:27   Like it's so much easier just to look at their blog posts

01:05:29   because it takes more time for the thing

01:05:30   and you gotta set it up and now you have the subscription

01:05:32   is automatically downloading and how does this podcast app work or whatever.

01:05:36   That's what I was talking about with the barriers to entry.

01:05:38   I don't think we're quite there yet.

01:05:41   And the other one might just be inherent that you can read much faster than you can listen

01:05:45   and people can talk much more than they can write.

01:05:47   Yeah, that's true.

01:05:48   I mean, that's always going to be a problem with audio and video.

01:05:52   The fixed time scale media is that they're much harder to skim than text.

01:05:58   so there's always going to be limitations there and inefficiencies and just restrictions

01:06:03   on how much time and attention people are willing to give to them.

01:06:08   Anyway, before we go on now, it's been so long before we go on to the Brent post. Our

01:06:15   second sponsor and our final sponsor of this episode is a returning sponsor. Once again,

01:06:19   welcome back, Hover as a sponsor. Hover is a domain registrar that doesn't suck, and

01:06:25   And I mean that from the bottom of my heart.

01:06:27   I use them myself.

01:06:28   They didn't tell me to say that.

01:06:29   I use them myself.

01:06:31   And I've used a lot of domain registrars.

01:06:35   I've never found another one that I wanted to stick with

01:06:38   after using them.

01:06:40   And Hover is the one I stick with.

01:06:42   They really are great.

01:06:43   I mean, first of all, go to hover.com/atp.

01:06:47   Let them know that you heard about this from here.

01:06:49   And you'll get-- oh, crap.

01:06:52   I forgot.

01:06:52   I think it's 10% off if you use that as your entry point.

01:06:56   Sorry about that, I don't have that in this script.

01:06:59   Go to hover.com/atp.

01:07:00   So basically, they have .net, .co, .com, .tv.

01:07:05   They have tons of different country code top level domains.

01:07:08   You can register with them.

01:07:10   They believe that everyone should

01:07:12   be able to take control of their online identity,

01:07:14   get your own domain, and they make it really easy to do so.

01:07:17   And I've advocated for years.

01:07:21   don't use like a gmail web address as your email address

01:07:24   don't use a google plus site as as your identity on the internet like you

01:07:28   by your own name and and do your own thing

01:07:31   make your identity they make it easy to do exactly that

01:07:34   and i mean just the fact that they're a regular star that doesn't suck that

01:07:36   should be enough reason right there that you should go try them out

01:07:39   uh... but occasionally more reason they have very very clean well-designed

01:07:42   control panel uh... it's very you know low hassle low friction

01:07:47   they even have this cool thing

01:07:49   where

01:07:50   They have, I think they call it like a migration service.

01:07:54   I forget.

01:07:55   It isn't in my list.

01:07:56   But they have a thing where if you're

01:07:59   willing to give them your login information for your old

01:08:04   registrar, they will go in there and migrate all your domain

01:08:07   names to their service for you.

01:08:09   Because transferring domains can be a bit of a hassle.

01:08:11   They will do all of it for you.

01:08:13   I think it's called domain valet service.

01:08:16   It's really, really cool.

01:08:18   I really do recommend them very, very much.

01:08:20   I mean, I can give you a whole bullet list

01:08:22   of what they told me to say here,

01:08:23   but they really enjoyed the last time

01:08:25   when I just told you how I feel about them,

01:08:27   and I'm doing most of that now

01:08:29   because they really are just that good.

01:08:31   Domain Rager stars, it's very, very hard to find a good one,

01:08:35   and they really are fantastic.

01:08:37   So thank you very much to hover.com.

01:08:40   Go to hover.com/atp, and this is run, by the way,

01:08:43   by the Two Cows people, which they've been in business

01:08:46   since the 1800s doing all sorts of geeky things.

01:08:49   So you can trust these people.

01:08:51   They really know what they're doing.

01:08:52   They've been around forever.

01:08:55   Everything rests on their reputation

01:08:56   that they've built for years.

01:08:58   So their customer service is fantastic.

01:09:01   They pick up the phone.

01:09:02   I mean, if you can imagine that, you call somebody

01:09:04   and they pick up the phone.

01:09:05   They actually do that.

01:09:06   That's real with them.

01:09:09   So yeah, go to hover.com.

01:09:11   They have email, forwards, DNS, all sorts of fun stuff.

01:09:15   Check him out, hover.com/atp, and thank you very much for supporting the show.

01:09:21   Yeah, so to add to that, just a couple days ago my father was looking to register a domain,

01:09:26   and he's very technically savvy for someone who doesn't do this sort of thing for a living.

01:09:30   But he doesn't know the ins and outs of DNS and those sorts of things, and so I was

01:09:34   like, "Dad, just go to Hover.

01:09:35   I have no time to deal with this right now.

01:09:36   See if you can figure it out, and if not, let me know."

01:09:39   And like 20 minutes later, he said, "Oh, I've got my domain.

01:09:41   It's all in Hover.

01:09:42   It worked great.

01:09:43   Don't even bother calling me back.

01:09:44   I'm all good."

01:09:45   So it's really--they're really good. So give them a shot.

01:09:49   Yeah, definitely. All right, so moving on.

01:09:52   So now I do still want to talk about this Brent Simmons article,

01:09:58   "30 Minutes to Sync." What he basically says,

01:10:02   he suggests a solution that, you know, iCloud in its current form for developers,

01:10:07   as we've talked about before in this show and as many people have talked about recently,

01:10:12   iCloud for developers has been really a mixed bag,

01:10:16   mostly bad. And any kind of like

01:10:20   database sync or any kind of complex sync for apps with iCloud

01:10:24   has been just tons and tons of problems from simply

01:10:28   being not functional to causing very difficult support problems

01:10:32   to weird bugs and the biggest thing being a

01:10:36   total lack of control over what's going on because you don't control iCloud as a developer.

01:10:40   And so what Brent suggests here is that Apple should make,

01:10:47   rather than doing the current quote magic setup where you

01:10:52   basically say, all right, here's my core data database,

01:10:55   sync it and hope nothing goes wrong, which doesn't actually

01:10:57   work in practice.

01:10:58   Rather than doing that, Brent says,

01:11:01   Apple should have basically a kind of thing

01:11:03   where you can define some server-side logic for them

01:11:07   to execute on your behalf.

01:11:10   And the app can talk to that logic, and most importantly,

01:11:16   and people can log into it, or you can log into it.

01:11:19   You can log into people's iCloud accounts

01:11:20   from your app and manage your own data, stuff like that.

01:11:26   And then they also would support things.

01:11:28   He said support periodic tasks like polling, Twitter,

01:11:31   or RSS feeds, or stuff like that.

01:11:33   And then he also points out, which we should point out also,

01:11:36   This is all very, very similar to a few systems

01:11:41   that already exist out there.

01:11:43   There's things like Heroku, which

01:11:44   is kind of like a generic app hosting platform that

01:11:48   is based on scalability automatically

01:11:49   and stuff like that.

01:11:50   There's also Windows Azure Mobile Services,

01:11:53   which we should disclose is a future sponsor of this show.

01:11:56   And Brent has actually done promotional videos for them.

01:11:58   And they do almost exactly what he's asking for,

01:12:02   but without the iCloud authentication and integration.

01:12:06   But I think, what do you guys think of this proposal?

01:12:09   Do you think it's realistic that-- first of all,

01:12:11   I think Apple never will do it.

01:12:13   But do you think this would be a good thing if they did it?

01:12:18   I think Apple is two degrees removed

01:12:20   from being able to pull off anything like this.

01:12:22   And if you think of the companies that do do it,

01:12:25   like Google and Amazon, the first stage

01:12:28   of getting to the point where you're even able to do this

01:12:31   is the company recognizes that one of the things they need

01:12:35   be really good at as a company is doing services in-house and that they can't do each one as

01:12:42   a kind of custom one-off special snowflake. What they need are infrastructure tools to

01:12:49   do that. So the reason Amazon builds up its infrastructure is we have to run Amazon.com

01:12:55   and it's actually a pretty popular website and we have scaling problems and we have these

01:12:59   problems that are in here, so let's build ourselves tools to do that. Once you do that,

01:13:03   like, OK, well, we've built these tools internally that are not great, but it lets us build amazon.com

01:13:11   or lets us build Google's stuff. If we change this just a little bit and shape it up, we

01:13:17   can sell this as EC2 or as a Google App Engine or whatever. That's the second part.

01:13:26   Apple doesn't even have that first part yet. I think they haven't yet recognized that internal

01:13:32   infrastructure tools are important enough to the company that they need to be as just

01:13:35   as important as like iOS or hardware design or silicon chips or whatever.

01:13:41   Maybe they're doing it internally, maybe they haven't figured it out, but that's what I

01:13:43   think is the main barrier to them doing this is they have nothing to bend because as anyone

01:13:48   who writes server-side software knows, you can't just say, "Oh, you guys, write some

01:13:52   code and we'll run it on our servers," because the first dude to put in an infinite loop

01:13:56   goes, "Oh, I guess we need some way to handle that.

01:13:58   Oh, I guess we need to handle resource usage.

01:14:00   usage. Oh, we're going to run your periodic job.

01:14:02   But I guess we need-- like, all the things that you deal with,

01:14:05   you have to just have an infrastructure for running

01:14:07   arbitrary jobs in a controlled environment

01:14:09   where you limit their CPU, but have error handling

01:14:12   so that you can let someone know why their thing didn't complete

01:14:14   because it was using too much of this.

01:14:16   I mean, even shared hosting providers

01:14:18   have to do crap like that where they're just trying

01:14:20   to divvy up their resources.

01:14:21   Apple is so not equipped to do that.

01:14:23   They're not even equipped to run their own things

01:14:25   in that manner.

01:14:25   In fact, they're farming a lot of it

01:14:27   out to run the Windows Azure stuff, which

01:14:30   great that Microsoft's good at that, but the fact that Apple isn't means that I would not

01:14:34   hold my breath for anything that Apple is going to provide where you give them code

01:14:37   and they run it on their servers for you. I would agree with all of that, and somebody

01:14:44   in the chat room is probably laughing by me saying that because I get a lot of grief for saying that.

01:14:47   But I agree with all that. Wait, whom do you agree with?

01:14:50   Both of you guys, really, that I think it would be awesome. I don't think Apple will ever do it.

01:14:55   I don't think Apple's capable of doing it right now.

01:14:59   But one thing I will say is that I wish that we had a-- this is going to sound so obvious,

01:15:05   but bear with me-- I wish that we had a little more insight-- or not insight, we had a little

01:15:10   more access to iCloud authentication, or iCloud in general, especially.

01:15:16   But one of the things I've been thinking about a lot in the spare time I don't have is to

01:15:22   write a very, very, very basic list, not to do, but like a grocery list app.

01:15:28   And what I mean by that is, and my wife and I, we use wonder list right now to organize

01:15:34   our shopping lists, one for the grocery store, one for hardware stores, et cetera, et cetera,

01:15:40   et cetera.

01:15:41   And although wonder list is very good, it's way more heavy handed than I need.

01:15:46   I just need a list of things that are either checked off or not.

01:15:50   One of the things I've been thinking about is writing my own, but the problem with that is I don't want to have to

01:15:54   in the case of Azure Mobile Services, my understanding is you can use Facebook to log in. Does this sound familiar at all?

01:16:00   You can use Facebook to log in, you can use Twitter to log in, you can use Windows Live to log in,

01:16:04   and I think there might be one or two other options or you can roll your own.

01:16:06   I don't have the patience to roll my own. I get no excitement and no jollies over rolling my own authentication and login system.

01:16:13   I wish I could just use

01:16:16   iCloud to handle that and I know I can when I'm talking about doing things in their own little silo

01:16:23   But I don't to my knowledge. There's no way for me to really

01:16:27   sync between

01:16:30   iCloud users easily and so it's stuff like this that Brent's talking about that I think would make it really cool and maybe make it

01:16:37   possible for me to make this grocery list app for my wife and I to share that uses iCloud as a login mechanism that

01:16:43   that I don't need to worry about the stuff I don't care about.

01:16:46   And Apple's, what's interesting to me is that Apple is usually much, usually very good at

01:16:52   abstracting away the things that I don't want to care about and properly abstracting them.

01:16:57   And in this case, they've, they've abstracted, as many developers have said, they've abstracted

01:17:02   too much and they haven't given, given us any, any vision into it. And it's just this big black box

01:17:07   that we have no control over and it stinks. And I really wish that a lot of these things that,

01:17:11   that Brent talks about were possible, but I agree with Jon. I don't think that Apple's in a position

01:17:18   that they're capable of delivering it, and I agree with you, Marco. I don't know that they have any

01:17:22   interest in it anyway. Well, they could do the limited version of this, which is they could,

01:17:28   like I just joked about earlier, there's no reason they couldn't provide a web service to authenticate

01:17:32   with your iCloud ID. Things that don't involve arbitrary developers giving Apple code that Apple

01:17:38   run on their servers on your behalf.

01:17:41   If they simply vended the services already available through the iCloud APIs, but did

01:17:45   it through an HTTP endpoint and allowed... because then you could write web apps against

01:17:50   it, right?

01:17:51   Which basically Apple does.

01:17:53   iCloud.com is taking advantage of its own...

01:17:56   I don't know if it is.

01:17:57   I know this is going back in time, but the iDisk website had a little JavaScript implementation

01:18:03   that would hit against a regular sort of web service endpoint to do its work.

01:18:08   And I'm assuming that iCloud stuff does similar things,

01:18:11   and it's not just all talking to Apple servers, which then talk to the services behind the scenes.

01:18:15   But there's no reason they can't expose some or all of the things

01:18:19   that you currently can only do from an iOS app as a bunch of web APIs.

01:18:24   Apple has never... I always say Apple doesn't understand the web.

01:18:28   I mean, that's kind of a harsh way to say it.

01:18:30   Apple doesn't want to understand the web because they want...

01:18:34   The reason all this stuff is restricted to iOS is they're not Google. They don't win

01:18:38   when more people use web applications or more people use the web.

01:18:42   They want people to use iOS devices, so there's not really a lot of incentive for them to say

01:18:46   "Oh, look, and now if you use iCloud as your

01:18:50   thing where you store stuff, you can make your web app hit it too."

01:18:54   Like, why would they want you to go to your web? They think you should be using an iOS application.

01:18:58   There's no real advantage for them to do that.

01:19:00   But that's one of the many, many other things

01:19:03   that keep people away from iCloud is it's locking you in.

01:19:05   You can't have a web app that uses that same back end.

01:19:08   You can't have an Android app that uses that same back end.

01:19:10   And certainly Apple doesn't want you to have an Android app

01:19:12   or anything else.

01:19:12   But in some degrees, there's no motivation for Apple

01:19:19   to do that.

01:19:19   But on the other side, they could put web service endpoints

01:19:24   for a limited set of things.

01:19:26   and that would make people less reticent to use it.

01:19:30   Why would I ever use iClevel?

01:19:32   People use it on iOS because like, hey, no sign in.

01:19:34   Your phone is already set up with your Apple ID.

01:19:36   It's great to not have that sign in thing, and it's perfect.

01:19:39   But if that's it, that's the extent of that relationship

01:19:43   with this authentication system, then it's like,

01:19:46   oh, but I also wanted to have a website.

01:19:48   And so now you gotta bend over backwards

01:19:49   and do something like Marco did, where like,

01:19:50   well, Newspan just works 'cause you're signed in

01:19:52   with your Apple ID and blah, blah, blah.

01:19:54   But I also wanna have a website,

01:19:55   but I can't get access to that ID,

01:19:57   so I gotta figure out some system,

01:19:58   and before you know it, you're emailing attachments

01:20:02   with weird data in them.

01:20:03   - I think also there's a problem here of,

01:20:08   I mean, first of all,

01:20:10   I guess I can see why people in our little world

01:20:16   of Mac nerds so often talk about this problem,

01:20:20   but as a developer of web services,

01:20:23   and again, because I have a history of doing these things myself,

01:20:27   and I've gotten pretty good at it recently, maybe I'm biased here,

01:20:31   or maybe my view is somehow disordered here, but

01:20:35   I don't see the big lineup

01:20:39   of people wanting to build web services on Apple's cloud infrastructure.

01:20:43   As a web service builder,

01:20:47   I would never want that, because they've shown that,

01:20:51   - They've shown partly that they don't really care

01:20:53   that much about reliability necessarily

01:20:56   or about improving it to the point where

01:20:58   their actions speak louder than words.

01:21:00   I'm sure the people working on it are trying very hard

01:21:04   to make it work, but it just seems like as a company

01:21:07   it isn't a very high priority.

01:21:08   And you can see that in the relatively little progress

01:21:12   it's made over the last few years,

01:21:14   it just doesn't seem like it's a number one priority

01:21:17   for the company.

01:21:18   - That's not neutral either.

01:21:19   way getting out a new iPhone is.

01:21:22   Like S3, you feel like, "Well, that's neutral." S3 is a bucket where I put crap. You understand

01:21:28   the business model behind it because you pay for usage, and you're pretty sure that Amazon—well,

01:21:34   maybe you're less sure these days—that Amazon is not—you're not in competition

01:21:38   with Amazon by making your application, your mobile application and your website that both

01:21:43   use S3.

01:21:45   But I think the other problem with Apple stuff is that because it doesn't seem like it's

01:21:50   that high of a priority for them all the time, it's gone through so much flux.

01:21:54   If I would have built Instapaper on Apple's web service of the time when I started it,

01:22:02   I would have been building it on something called .Mac, which most people have probably

01:22:08   never even heard of these days.

01:22:09   Most people who use it have probably forgotten about it.

01:22:14   First it was .Mac, well first it was whatever it was,

01:22:17   like NetTools or whatever.

01:22:19   - iTools, not Netcast. - iTools, that's it.

01:22:20   - iTools, yes.

01:22:22   But there was no services.

01:22:23   - Internet blast, something.

01:22:25   And there was that, and then there was .Mac,

01:22:29   which had some internet service-based things,

01:22:33   and then MobileMe, which had more

01:22:35   internet-based service things,

01:22:37   and then iCloud, which has actually fewer

01:22:39   internet-based service things.

01:22:41   And there is, like Apple's own system is in such flux here,

01:22:46   I would never want to build on that infrastructure.

01:22:50   - That's exactly what I was gonna ask you.

01:22:52   - Well if you think about, but if you re-conceptualize

01:22:55   all those things you just listed, including iCloud,

01:22:57   the current one, not as infrastructure,

01:23:01   which clearly Apple doesn't think of them as,

01:23:04   but merely as ways to make it easier

01:23:08   to develop feature-full applications for their platform.

01:23:13   So, you know, if you look at it that way,

01:23:16   MobileMe, .Mac sync and all that stuff

01:23:18   was a way for you to write a Mac application

01:23:20   so that you didn't have to write the sync stuff.

01:23:22   Game Center, so that you can write a game

01:23:24   so that you don't have to do the matchmaking,

01:23:26   you don't have to do the server-based scoring

01:23:28   and stuff like that.

01:23:29   'Cause maybe you just write iOS apps,

01:23:30   maybe you just write Mac apps,

01:23:31   maybe you don't know or don't care

01:23:33   or don't want to do the server-side stuff.

01:23:34   So they're trying to make it so, hey,

01:23:36   you want to write a game, but you don't want to maintain servers,

01:23:41   write it against Game Center.

01:23:42   You want to write a Mac application,

01:23:43   but you don't know anything about syncing,

01:23:44   well, we have a framework.

01:23:45   Same way any framework works.

01:23:47   Frameworks do the work for you so that all you need to know about

01:23:50   is dentistry for your dentist app, plus a little bit of cocoa

01:23:53   to do the UI.

01:23:54   But you don't need to know about OpenGL to make OpenGL animated stuff.

01:23:58   I was going to say layer kit.

01:24:00   Core animation does that for you, right?

01:24:03   If you look at iCloud in that respect, it's like, well,

01:24:05   A, why the heck would it ever work outside of iOS?

01:24:07   Because the whole point of it is just

01:24:08   so that you can make a cool iOS app that

01:24:10   has lots of interesting features without having

01:24:12   to write the server stuff.

01:24:13   And I don't have a B. Or maybe I had a B and I lost it.

01:24:18   I don't know.

01:24:19   But if you look at it that way, it looks less insane.

01:24:23   But first of all, if it doesn't work well and whatever,

01:24:28   then as we've talked about at length at other shows,

01:24:30   even in that respect, it doesn't work.

01:24:32   So iCloud is a failure, even if you narrow it

01:24:35   just to that focus.

01:24:36   But the reason we're talking about broadening it out

01:24:40   to having web services and stuff like that is we

01:24:44   want it to be an infrastructure component, because lots

01:24:46   of people who write applications don't just

01:24:48   want it to be easier to make a cool iOS application.

01:24:51   They want it to be easier for them

01:24:53   to run a business or something.

01:24:55   So S3 provides that, because I don't want to deal with storage.

01:24:58   But if I pay S3 for my usage, I get storage.

01:25:01   And that solves the storage problem

01:25:03   for my business product provided people buy my product,

01:25:05   or I have some other revenue stream that I can put towards that.

01:25:09   Whereas the iCloud thing, it's like, well, OK, that solves my problem

01:25:12   how to write the iOS app.

01:25:13   But OK, now when I go to write the Android app,

01:25:15   I have to do something different.

01:25:16   And now when I write the web app-- maybe if Google does it well,

01:25:19   you can write the Android app and the web app

01:25:20   against the same infrastructure.

01:25:22   And Apple looks like the odd man out of like, oh, I

01:25:24   got to use iCloud with that.

01:25:25   Or maybe I don't.

01:25:25   Maybe I'll just shut an iCloud like everybody else

01:25:27   and just use whatever the other infrastructure thing is.

01:25:30   And then once again, it ends up with Apple being fenced out

01:25:33   the infrastructure business. Because like Marco said, why would I ever trust them for infrastructure?

01:25:37   Because it doesn't seem like that's what they're making. They're making this other thing.

01:25:40   Yeah, but like I said earlier, as you just said, I think Apple's

01:25:44   Apple's technologies, their developer technologies, are so good at abstracting away the things we don't care about. A perfect example is what you brought up with core animation.

01:25:52   I mean, I don't know crap about

01:25:54   anything

01:25:56   related to animations or OpenGL or anything like that,

01:25:58   but I can animate the crud out of a UI view like the best of them because the

01:26:04   coronation API is so simple and so robust. And it's really disappointing that Apple

01:26:10   doesn't have an equivalent, just like you said, for what we call cloud computing.

01:26:15   And I don't think I would use it even if it was available tomorrow, but man would I be,

01:26:20   like say at WWDC, if they announced some sort of cloud computing platform, I would be really

01:26:26   really darn excited about it. I surely wouldn't be the first to use it, but I'd be really

01:26:30   excited about it.

01:26:32   I would stay very, very far away from that platform, honestly.

01:26:34   Well, I mean, it's also about commitment, because the fact that they have turnover—if

01:26:39   you think of MobileMe and Sync as like, "Well, that was the thing we had five years ago,

01:26:44   and you wrote and sold your application five years ago, and you made your money from selling

01:26:48   that application, and that time is over, and now we assume you're going to write a new

01:26:51   application," and people like—bare bones with Yojimba—are like, "Well, actually,

01:26:54   still got the old application and we'd still like to keep selling it because we kept updating

01:26:58   it." And Apple's like, "Yeah, no, that time has passed. We have a new thing now. So either

01:27:02   rewrite your application against iCloud, which doesn't seem to work even as well as Think

01:27:06   Services did, or come up with another solution." Whereas if you think of something as infrastructure,

01:27:12   like if Amazon decided S3 was our first attempt at storage, but we're going to sunset S3 and

01:27:18   we're going to replace it with S4 because that's one more, right? And it's totally incompatible.

01:27:23   People will really regret ever starting a project that relied on S3 as the backend.

01:27:27   Like I don't know, Dropbox, a couple of small companies that might be impacted by this decision.

01:27:32   And once you get that reputation of like, "Oh, they're not interested in building infrastructure

01:27:37   for long-term support," and especially with things like web services and stuff where you're

01:27:41   not touching it, where it's all over network connections, there's really no reason ever

01:27:45   to completely sunset that type of thing.

01:27:49   You can change how S3 works, and I'm sure it has changed behind the scenes.

01:27:52   all the hardware, change all the software that runs it.

01:27:54   You just have to present the same API endpoints

01:27:56   to the network, and then you're good to go.

01:27:59   But Apple, it's not making infrastructure.

01:28:01   There's no pretense of making infrastructure.

01:28:03   It's changed 17 times since then.

01:28:04   So that's why everyone is living in deadly fear.

01:28:06   So if Apple ever did come out with something like this,

01:28:09   they would have to explain, say, upfront,

01:28:11   we understand the difference between infrastructure

01:28:14   and the thing that makes it easier for you

01:28:15   to write an application that you're

01:28:16   going to sell for three years.

01:28:18   And this is not that other thing, so trust us.

01:28:21   and then still they would have to earn that trust and do a good job and so on and so forth.

01:28:26   But there's a big gap.

01:28:28   It just seems like Apple wants to work at a higher conceptual level than that. They

01:28:34   don't want to expose the guts. They don't want to expose server-side functionality.

01:28:38   They don't want to expose low-level APIs or give you low-level data access like that.

01:28:44   They want it to be the Ubiquiti iCloud API that we expose like this in the apps, and

01:28:51   that's it.

01:28:53   I think that's much to their downfall, really.

01:28:57   There's a few cases that benefit from that, but there's so many that don't.

01:29:02   As we've discussed in one of the earlier episodes of this show, I think the whole concept of

01:29:08   that is flawed.

01:29:09   way they've tried to do like iCloud is a simple thing. Just hit this flag and call this function

01:29:15   and it works. And it's not that simple. It doesn't just work. And the whole concept,

01:29:23   I think, is flawed. But I don't see Apple ever going lower down the stack and ever offering

01:29:29   that kind of deep, low-level access, execution of arbitrary code in any language with any

01:29:37   I don't see that happening.

01:29:39   Any kind of server-side stuff they would do

01:29:43   would probably be in the form of things that you initiate locally

01:29:49   in the apps that maybe it would be like Newsstand API,

01:29:53   where you could enqueue a background download

01:29:57   with a certain push notification or something like that.

01:30:00   It would be these restricted cookie cutter templates.

01:30:03   It wouldn't be like, "Oh, you can execute arbitrary logic on the server."

01:30:06   And I just don't see them ever getting into that business.

01:30:09   Well, the key value store, we keep saying iCloud, but we're really mostly talking about the core data thing,

01:30:14   but key value store is an example of limited, simple functionality that does not dictate implementation.

01:30:20   It's just so super limited, like it's for preferences and stuff, it's not for like, "Hey, store all your data here."

01:30:25   But you can imagine that, there's a difference between SOAP and WSDL and all that versus REST.

01:30:32   like you can imagine a really simple HTTP API for key value

01:30:36   store, right?

01:30:37   It's like it writes itself.

01:30:38   It's so simple because the API is so simple.

01:30:41   And no one is complaining that key value--

01:30:42   Because the concept is simple.

01:30:43   Yeah.

01:30:44   And no one is complaining that key value store shouldn't

01:30:45   exist, right?

01:30:46   But what Brenda's talking about is that kind of thing

01:30:49   makes sense because they're just having infrastructure.

01:30:51   So let me, as the application developer,

01:30:53   worry about all these other details.

01:30:54   I just need a place that's not here.

01:30:56   I need a place that's across the network for me to put my stuff

01:30:59   and to be able to get it back later.

01:31:01   I'll take care of the other details.

01:31:02   Like, it's the bargaining stage.

01:31:04   It's like, you know, I know you want to do everything for me

01:31:06   and make it magically work, but that's not working for us.

01:31:08   So if you just gave me a place to put stuff

01:31:10   and I could still use my iCloud authentication, which

01:31:12   is really great, I'll take care of the rest of it,

01:31:14   and then I'll have a good application.

01:31:17   Which it's stepping down.

01:31:19   It's saying, I'm willing to do more work

01:31:21   to talk to your dumber bucket if I can authenticate

01:31:24   against that bucket using the iCloud ID that's already set up.

01:31:27   Please, Apple, at least just let me do that.

01:31:30   Alright, and with that we should wrap up.

01:31:33   Thank you very much to our two sponsors, macminivault.com/try.

01:31:38   You can go co-locate your own Mac Mini for $30 a month and less for the first three months

01:31:42   if you use coupon code ATP50 for 50% off.

01:31:46   And our second sponsor was Hover.

01:31:48   Go to hover.com/ATP, get a nice discount there, register domains in a way that doesn't suck.

01:31:54   They are the best registrar I've ever used and they're awesome.

01:31:58   Go to hover to register all your domains and related domain activities such as email, forwarding,

01:32:02   DNS, et cetera.

01:32:04   Thanks a lot to our sponsors and thank you John and Casey.

01:32:10   Now the show is over.

01:32:12   They didn't even mean to begin.

01:32:15   Because it was accidental.

01:32:16   Accidental.

01:32:17   Oh, it was accidental.

01:32:19   Accidental.

01:32:20   John didn't do any research.

01:32:22   Marco and Casey wouldn't let him.

01:32:25   Because it was accidental.

01:32:27   Accidental

01:32:28   It was accidental

01:32:29   Accidental

01:32:30   And you can find the show notes at ATP.fm

01:32:36   And if you're into Twitter, you can follow them at

01:32:41   C-A-S-E-Y-L-I-S-S

01:32:45   So that's Kasey Liss M-A-R-C-O-A-R-M

01:32:49   N-T-M-A-R-C-O-R-M-E-D

01:32:51   S-I-R-A-C-U-S-A-C-R-A-C-U-S

01:32:56   It's accidental, accidental, but you didn't mean to.

01:33:02   Accidental, accidental, tech podcast so long.

01:33:10   Thanks.

01:33:11   That would have been a tight show

01:33:12   if you hadn't had the middle section for half an hour

01:33:15   where we talked about podcasts on a podcast.

01:33:19   That you derailed yourself.

01:33:20   Like, you did that to yourself.

01:33:22   I kept meaning to make that just like a blog post,

01:33:24   and I haven't had time.

01:33:26   So this happened.

01:33:27   So now we all get to pay the price?

01:33:28   Yes, exactly.

01:33:29   You can still make it a blog post.

01:33:30   That's the grand tradition.

01:33:31   I think it was interesting.

01:33:32   Actually, I thought it was interesting too.

01:33:33   It is, but you're the one who's always against talking about podcasts on podcasts.

01:33:37   I don't have a problem with it.

01:33:38   I think it's fine.

01:33:39   Well, no, I think it's okay to talk about podcasts on podcasts as long as you make fun

01:33:42   of it first.

01:33:43   All right.

01:33:44   Then it's okay.

01:33:45   Then it's okay.

01:33:45   then it's okay.