Transcript

00:06 [Music]
00:13 Uh, we get to see a a second look, right? Funny funny slide. A second look,
00:20 our first second look at stack six, right? Um, Isaiah, can are you uh can I
00:28 hear you? Oh, looks like you're muted, bud. Yeah. How's that? There we go.
00:37 Hey, everybody. Wait, can everybody uh hear me? Okay. For the event and everything. Nice.
00:43 What's that? I Oh, you shaved for the event and everything. Look at that. I clean up, you know.
00:53 Try to look nice for you, Joe. Um, uh, yeah.
00:58 Uh last year um I did a a demo of stacks and I mean it went pretty well but
01:04 um it was a lot of talking and slides and not much demo. Um and then I did a second talk where it was just talking.
01:14 Um so this year I'm going to try to do mostly demo and just a little bit of slides. Um so
01:23 uh yeah let's let's get to it. So, um it's been about a year since most people have seen uh an update of stacks.
01:34 You know, I I keep a a small group of people um uh appraised of the situation, but um yeah, I've been keeping it close
01:43 to my chest for a while. So, um I'm going to share some stuff that is like just on the edge
01:52 of functional, some stuff that's really rock solid. Um and then I'm going to give you guys a demo uh that you can
02:12 But hopefully this weekend, I think it it's it's an easy bug to fix, but I just didn't want to release it while heading
02:21 into a presentation. So, um, but we are going to use that version here. So, we'll we'll see it. Um, so what's today going to be about? Uh, let's see.
02:32 Um we're going to start with a quick demo and we're going to do this kind of in layers. Uh the demo first we're just going to compare and contrast this year
02:41 with last year. Um and then I'm going to kind of peel back a layer and we're going to look at a few new pieces of
02:49 user interface. And this is mostly uh some basic stuff. Um but it's brand new stuff. Um so we're going to talk talk
02:57 about the open quickly interface. um the new link window which kind of uses the open quickly interface um about deploy
03:05 so FTP and um about the project window.
03:11 So yeah uh that's just some new user interface that's fun to see. Um I guess
03:17 FTP is is really uh a totally new feature um that you know is more than
03:25 just surface level. Um, and then, uh, going to peel back one more layer and we're going to look at loading the
03:34 stacks library. And I have to admit, I just changed this. Um, I had a totally separate third thing planned uh, for
03:43 couple weeks. Um, and I ripped it out like two days ago. So, um, this could go really well or really badly, but um, I
03:52 think it's going to go really well. Um, I'm optimistic. Um, I'm going to talk about the progress that we've made on
04:10 there are good reasons for that, but um, I knew we could do better. And so last year I got to the conference and I said,
04:19 "Hey, um I've worked all this time on the uh loading stacks interface and it's it's
04:27 so much better." And then I heard somebody say, "Yeah, it's about the same." And I was so disappointed that I
04:35 hadn't, you know, really made real progress um that when people put it actually to the test, they couldn't even
04:42 tell. Um, and that was, yeah, I felt, uh, embarrassed for myself that I had done that. So, this year we're going to do better. I promise. All right.
04:54 Um, so without further ado, let's take a look at uh just a surface level demo.
05:01 So, um, yeah, we're going to review and compare uh last year with this year. So,
05:09 let me see if I can uh undo this and Oh, yeah. I should have done that.
05:17 Well, zoom that.
05:19 And play that. And we'll just zoom over here. Here.
05:26 No. Here. Here we go. All right. So, um you turned off your screen share.
05:34 Oh, really? Did it? Yes. Oh, because I switched desktops. Huh. Is it back or is it just totally gone? No, it's it's
05:42 totally gone. You have to hit the share again.
05:45 Great. All right. Um well, we'll do it on the screen then.
05:50 Um give me a sec.
05:59 Share. Um, one one none of these look right. How about
06:06 this one? I think before you were sharing just your keynote and when you closed Keynote, it uh it it stopped the share, I think. Oh, interesting. Okay.
06:15 Well, maybe this will be better then.
06:17 We'll have to see. I I kind of have a feeling it's like only sharing and I'm I'm see um but we'll uh we'll try this.
06:28 Um, I need to get this out of the way.
06:32 Uh, so here are the two versions of stacks that I'm going to kind of look at and compare. Um, this one is from just a
06:41 few days ago. Um, and this one is from last year. So,
06:49 um, we'll hide that one for a sec and let's just take a look at where we were last year.
06:58 Um so I think the first thing to notice um about
07:06 uh last year's is that a lot of the tools at the top here in this this tabbed interface um have changed a
07:15 little and progressed a lot. So um one bit of feedback I got from yeah a
07:22 handful of users at least was that um whenever you selected a different thing so I'll I'll select over here but watch
07:30 over here it kind of jumps around and if I select page it jumps to a new set of tabs
07:37 um I did that because uh that is the way Xcode works um and I
07:45 was used to it And it's moderately understandable. I mean, it is kind of complex, but it's pro app, you know. Um,
07:55 but I got a lot of feedback that that was kind of weird and disconcerting and they didn't know what was going on all the time. Um,
08:03 so I took a step back and I looked at what the common pieces of the common possible tabs there were.
08:14 um and try to fit things in to a a framework where um it doesn't have to jump around
08:21 anymore. So um let's take a look at uh this here. So this guy, if I select different
08:30 items, um you'll notice it stays in the same tab no matter what. So, um, if I
08:38 select a stack or a page or a library item, um, it's never jumping around over
08:45 here. If there's nothing to show, you get nothing. But, um, at least it's not kind of confusingly, uh, bouncing around.
08:56 Um, I think there's still room perhaps someday to make some of the items um
09:04 maybe come and go or disable uh when necessary, but um I think for a first version to let
09:12 people get used to the idea um keeping the tab stable is just a good
09:20 thing. Um what else is new? Um let's take a look at this new version. Uh we've we've got a few things that came
09:28 back. Um the library uh library groups are back. Um we have
09:34 stack updating is back. Uh we have uh deploy and a deploy settings buddy button. Um the site settings are back.
09:44 Uh and now there are um yeah both site settings and uh
09:53 settings for the app itself. So um you know some things uh you want to change in every project. Some things you want
10:07 have uh a new tab just for library items that shows uh details about the library item that's selected. If you do things
10:15 like doubleclick on a library item, um even if something else is selected, it will jump to that item. So, there's some nice little shorthands and shortcuts.
10:25 Um, we have uh a a theme column to the library now that looks a bit different
10:32 and shows the site theme that's selected. Um, and uh there's a lot of
10:40 other little details. Um, for instance, if you have um a theme that has a lot of settings, uh you'll see those over here.
10:49 Um and uh you can then uh modify those settings by override um in the page
10:58 column. So there's just a lot more. Um whereas last year there were a few things that were just disabled. I was
11:06 like, you know, that's not stable enough to show everybody. So I'm just going to disable the button and move on. Um so a
11:14 lot more is is filled out. In fact, there's only a few things that are Yeah.
11:19 grayed out anymore. All right. Um, so that is just like a quick overview to check out on
11:28 the surface level what has changed, but let's um yeah, peel back an onion layer
11:36 and go to the next level of detail. So, um, okay. Here, I'm going to switch
11:42 desktops and see if I can go back to Keynote. Did I lose my sharing? Nope, you're good. Hey, we rock. All right.
11:53 Um, so I'm going to put this on my other screen
11:59 and hit the play button here. Play. Why not
12:06 play? What's going on? escape. Is it Is it already open somewhere else? Is it full screen somewhere?
12:15 It shouldn't be. No, it's like it's locked up here.
12:20 Keynote looks grayed out. Oh, there it went.
12:23 Did I lose my sharing? No. No, you're still there.
12:28 Um I'm just going to open it up again cuz Oh, wow. I hope it didn't like delete my
12:36 presentation. That would really bite. State of the stacks
12:45 2025. All right, I think we're good. Sorry for the mixup. Um, I only do
12:55 this presentation stuff once a year, so I have to relearn how to do it every year. Um, okay.
13:02 Uh, so what's new? Uh, let's talk about just a few more things.
13:07 um open quickly the link window deployment um and the new project
13:14 window. Um but I wanted to mention just while we're here that there were a lot of other
13:22 little bug fixes. This is not all that we worked on for a whole year. Um, and uh, I went back to my change list and I
13:32 just kind of copied into into a few slides so you could see.
13:36 Um, yeah, we had a bunch of fixes and then we had a bunch more and then we had
13:45 more and then we had more. Um, and then I got tired of making these slides. Um,
13:51 so that only goes back to version 20. 10 was the last public version. So, um yeah, that's a wait, that's twothirds.
14:03 Um there were a lot of bug fixes. That's all I have to say. Uh and uh bunch of new little features. Um some are small,
14:10 some are huge. Um but there's a lot in there and uh the app is becoming much more real. Um not so beta
14:20 anymore. Okay, so let's actually do the demo here. Um, I'm going to switch back and we're going to hope for the best.
14:29 Um, going to minimize that. Right. Okay. So, uh, the first thing I want to talk about is, um, open
14:38 quickly. So, um, in a lot of pro apps, uh, that have,
14:46 you know, projectbased stuff where there's a zillion things to click on in your project,
14:53 um, often times, you know, digging through those things can be really kind of cumbersome
15:03 um, to go look for something. And you might even get so many pages in your project and layouts in your project or
15:11 whatever that um you don't even remember where you put something. You know it's in there somewhere. Uh maybe you remember the
15:19 name of it or something that's uh associated with the page and you just want to go find it. So um now uh we can
15:27 choose you can either choose it from the the file menu um which is open quickly.
15:34 It's right here. Or you can do command shift O which um is the same as the shortcut in I think omni graffle has it
15:43 or omni node. Uh one of the omni apps has it and uh xcode has it. A bunch of apps have it. Anyway um command shift O
15:52 you get a little sherlock uh spotlight like um uh window and you just start typing. So,
16:01 I noticed that there was a page in there called page six. So, if I just type six, um, that's what comes up. You know, as
16:08 I'm typing, it is showing me the things that are applicative there. So, I type six, it jumps to six, and I'm good to
16:17 go. Um, it becomes such second nature using this feature
16:24 that I barely even use the sidebar anymore. And if you're on a small screen, I work on a MacBook Air a lot.
16:31 Um I've grown to really love my MacBook Air. Uh the screen is small, so um you
16:38 know, you you can uh get rid of the sidebar and just, you know, open whatever page you need to um and jump to
16:48 that page. No big deal. Uh I noticed also that there was a page in there
16:55 called page five. So, I can jump to that. Um, let's shift O five. Boom.
17:03 Um, in here, uh, I'm going to change the folder name to, I don't know, how about
17:12 42 since everything is named with numbers.
17:18 Um, and now if I go back to yeah my homepage say um and I do command shift O
17:26 I can type uh 42 and it knows that too. So there's more context built in to the pages and
17:36 it's indexing all of that content when you open this window. Um, so now we can jump right to that page. So you don't
17:43 even necessarily have to remember uh yeah that it was
17:51 the the actual page name. Um other things that are associated with the page
17:57 um are also indexed. Um I noticed uh we're using version 33 here. I need
18:06 35 to show you the next feature. So, um I'm going to Well, I'm just I see that is 35. No.
18:15 Oh, is it? Did I just in the in the in the top in edit mode? It says 35. Oh,
18:22 thanks, Joe. Um the the document is called 33, and that threw me off. I I made one document for this demo, and I
18:30 it was for version 33. I've since built version 34 and 35. Um and uh 35 has the
18:38 bug fix in it. So it's for the link window. Uh so let's look at the link window. Um I have a button here. Uh
18:45 everybody's familiar with uh buttons. Um and yeah, so I'm going to select the
18:52 button and then which looks really weird. For some reason I have turned off
18:59 the Oh, it's probably because I'm on a weird Yeah. Okay.
19:05 Um yeah uh themes can apply styles to edit mode.
19:13 Um this is my reason variant that um I'm going to use as a
19:22 demo in tomorrow's talk about developing themes. Um okay. So
19:30 uh select the button. Let's take a look at um the settings for this button. And
19:37 let's check this edit link here. In fact, let's start with a clean button because that one already had a link in
19:44 it. Uh boom. All right. Clean button, no link. Uh I click edit link. So, uh this
19:53 is a lot like the open quickly uh interface. Um I noticed one of the things that I have
20:01 always yeah felt is cumbersome in like every web development tool I've used um
20:08 is that whenever you want to make an internal link to the in inside the document to another page or to an asset
20:16 or whatever um you have to go hunt for it in a big list or a big menu and it's always a pain.
20:24 Um, and hey, why not use the open quickly interface to go find what you need in
20:31 uh for links? So, I can do that same thing here. I can link to page six and um yeah, lock it in and it shows me uh
20:40 details about it here. So, it's a little bit different than open quickly cuz um it doesn't immediately like lock in.
20:48 There's more detail to links. So, yeah, it links to page six. Um but if you need to edit properties, you can also do that. Okay.
20:58 Um yeah, then we hit link and we get a link. Um so yay link. Um and
21:07 uh it also works um for yeah just any link you want to link
21:15 to out on the internet. So instead of typing a, you know, a word that's associated with your document, you just
21:22 uh type the link. So let's try um yeah, it has a lot of uh built-in links. Like if I just start typing
21:30 Apple, it'll link to Apple. Um but let's link to something more interesting. Um, so let's
21:39 go to like uh let's go to my website, but I'm going to go to the stacks page
21:48 itself. Okay, so um yeah, I I lock that in and uh it shows me the link of where it's going to
21:56 go. It's going to go to stacks and I hit link and boom, I I have a link. Um nothing surprising there really.
22:05 Uh, it's actually pretty trivial. I mean, that's the whole point is um I wanted to make building a link as like
22:13 seamless and um simple as possible cuz it seemed overly complicated um in other languages
22:22 and other platforms. So, um but I want to do one more thing.
22:29 Let's go um over to page six where I put the link uh sent the link and I'm going to put a button in
22:35 here and I'm going to click edit link and now when I made the link to stacks
22:43 before to the stacks page before um I had to type in the whole thing but now I'm just going to type in something
22:51 that's part of that link. So, I'm just going to type in stacks. Um, and it knows
22:57 that that link is somewhere else in the project, and it's already indexed it and
23:04 added it to the open or the link quickly uh menu, and I can just lock it in. And
23:11 now, uh, those two links are are the same. Um, even if you set interesting properties on that other link, you'll
23:20 get them here, too. Um, you don't have to have them here, but yeah, they come for free. Um, so again, uh, I'm just
23:27 looking to reduce the amount of work necessary to do some of the basics. Um, I was
23:35 already having to build a a link window cuz that's, you know, the plugin, something that I I rely on Rapid Weaver
23:43 for. So here I had to build it. I just decided, let's make it better. Um,
23:50 right. So that is Oh man, I'm already running out of time. Um, so I need to move a little bit quicker. Remember, we started late, too, so don't worry about
23:58 it. Okay. Well, I I still try to stick to about an hour. Um, but uh yeah, if I need to run a little bit longer on the
24:06 last demo, I I can. Um, so uh yeah. Uh what am I what? Uh oh,
24:16 deploy. Yeah. So, hey, let's check out uh FTP just quick. We'll we can breeze through this one. Um it's it's I mean,
24:23 you know how FTP works. Um so, I got a deploy settings. Um and yeah. Uh oh, no.
24:31 This pro this file doesn't have the stuff.
24:35 Um do I have And I don't have that stack either. Damn it. Um all right. Yeah, I'm
24:42 not going to type in my password in front of you guys.
24:46 Um, do I have the other project somewhere? It might be in this
24:57 one. Yeah, that has that in it. I think it probably does. Nice. Okay, better. Um,
25:06 so yeah, it's it's uh FTP. In fact, it's it's SFTTP.
25:12 Um, I made the kind of command decision that for this version, um, we're we're supporting one type of FTP and it's only
25:20 the secure one. Um, if your hosting company doesn't support secure FTP, you need a new hosting company.
25:31 That's just all there is to it. Um, there is also FTPS. Um,
25:38 FTPS is also sort of secure. Um, the commands aren't sent over UDP anymore
25:45 and password isn't in the clear. Those are all good things. Um, but it's kind of quirky and it's much slower than
25:53 SFTP. SFTP is just built on the standard secure shell. Um, so it uses the same
26:09 honestly um at this point um and I do have the power to bring back
26:16 old FTP if it's like really essential for something, but I I really don't think we should. Um I I think it's a
26:24 disservice Excuse me. My mouth is getting too dry. Um, I
26:31 think it's a disservice to users to um, present them with something that is
26:37 known to be insecure at this point with the internet the way it is. Um, yeah, we
26:45 got to get away from the old crap, frankly. Uh, so we're doing uh, SFTP.
26:53 Uh, we also can do maybe in the future.
26:56 Uh so version seven say uh we can do uh SCP um that's allows us to like bundle
27:06 everything up in a zip file. SCP1 file SCP is secure copy um to the host and
27:15 then unzip it on the host. Um that's much faster actually. Uh, but it only
27:22 works for higherend hosts like VPS's and and you have to have shell access.
27:28 And I know that's probably um fine for some users, especially for
27:36 the pros that are building sites for other people, but it might be um yeah, it's just a bit advanced. So, I'm I'm keeping it simple for this one.
27:46 So, let's check out how it works. Um, yeah, it's a it's FTP. You you put in your host name. I use my IP address.
27:55 Um, you give it a folder where you want the stuff to go. Um, your your
28:02 uh username, password, and then uh I got a nice little validate button. Um, this I stole the idea right from uh Panic
28:11 Nova cuz I always like that they have a nice validate interface cuz it's always a pain. Anyway, so uh you hit validate.
28:22 Um and you get a check mark if you did everything right. Um but if you didn't, uh let's just see what happens. So I'm
28:29 going to delete uh the last R off my path name. And so now even if the username and password
28:37 is acceptable, some part of this setup is not going to work and it's going to break all of your FTP. So we'll hit the
28:46 validate again. And this time um it comes up with a pops up a mini log immediately and it
28:54 says ah you you failed something. So you check check check check check check check down at the very end it says the
29:01 path was uh not a readable directory. So it's not a real directory and it's not going to work. So um just you know add
29:10 the R back, hit validate and hopefully it'll work again. I think I might have just crashed things.
29:22 Yeah, sorry about that. It happens.
29:26 Um, let's uh let's kill this guy. Force quit. And we'll just quit that one too
29:35 while we're here. And I think I have a window here. I can just
29:44 open that. All right. Let's try again, shall we? Um, will it validate? Good.
29:56 Uh we'll cause a problem. Validate bad.
30:02 Um try again. Fix problem. Validate. Good. Hey, it works.
30:11 Um sorry, crashes happen. It's uh nature of the beast when you're working with beta software.
30:18 Um I it gives me a chance to talk about um so since last year I have added a crash reporter built in. So um if the
30:28 app does crash uh I get immediate feedback um through Sentry um and Sentry
30:35 is a yeah a lot of people are using it now.
30:39 It's kind of the best crash reporter going today for Mac apps.
30:45 Um, and yeah, and some of the nice things about it, it's very standardized. They
30:52 take care of security and stuff. So, um, when you have a crash report, there's the potential that like your IP address
31:08 stuff. I don't know who crashed it or or why. Um, maybe someday we'll add like a feedback part of it too. I haven't done
31:16 that part, but um, for now uh I don't have to beg for people to send me crash reports anymore, which is kind of
31:24 awesome. Um, cuz I did have to back in the plug-in days. Um, so I I just want to mention too, um, when you see the
31:31 check mark, you can click there and still see the same log again and you can see that it all works too, just in case.
31:38 Um, we'll say okay since we uh twiddled with things. And then you can hit the deploy button over here. Um, you get a
31:46 little deploy window. We're keeping it pretty simple in this version. Um, and yeah.
31:54 Uh, we'll hit deploy. It's going to validate again. Uh, just make sure before it starts trying to publish and
32:01 you copy and all that stuff. So, um, you'll see the validation happen. It'll export the site. This site is really
32:08 trivial, so it exports pretty fast. Um, and then it's going to update the file.
32:13 Uh, I'm not including in this version the comparison and partial upload. So,
32:22 it's just uploading everything, which I know is kind of a drag. Um, for a big site, it really does take a while, but
32:30 um, that's something I've been working on just recently. So, um, won't be long.
32:36 All right, that's deploy. Um, yeah, seven minutes. I have seven minutes to cover two whole topics, one of which was
32:44 supposed to be like 20 minutes long. So, um, I'm going to go over a little bit. I I'm really sorry to the people that are in Europe. Um, uh, I hope I'm making it
32:53 worth your while. All right. Um, yeah, we can do this last one
33:08 where it doesn't open your last open file, this is what you see. Uh, boom,
33:16 boom. You get a nice little project window. Um, and I want to give a shout
33:22 out to Dubox, Gary, uh, from Doobox who designed this for me in Swift UI, uh,
33:29 and sent me, yeah, it was like fully functional basically. Um, he's a Swift UI master. So I can't use the Swift UI
33:37 version unfortunately because that would constrain us to just Soma and uh and
33:46 Sierra Mac OS versions and I I do want to support back to Bixer at least and maybe even one or two versions after
33:53 that before that. Um so that means I can only use just like the most basic Swift
34:02 UI stuff. Um, so even something as simple as this is not possible unfortunately. Um, because Swift UI is
34:10 kind of fun. Um, it's also kind of slow, but whatever.
34:15 Um, so yeah, it's a project window. You know how how they work. Um, you probably have seen them in a zillion different
34:22 uh, project- based apps. Um, you can select something, you press return, and it opens the file. That's it. Um, if you
34:32 don't have a a file open, it auto opens this window. Um, and if you
34:38 launch the app um, oh, it did it on this screen. Uh, it will, um, pop open the
34:47 window automatically. So, you know, you can see your recent documents um or you can open a new document or create a document in one kind of easy step.
35:08 with that bit and we can can I
35:17 boom. Hey, let's talk about performance. So loading the stack library that's been a drag, right?
35:27 Um and like I said last year at the conference I I think offhand mentioned
35:34 that I had improved the performance and I thought substantially.
35:41 Um turns out not so much.
35:45 Um you know after that I started doing some real performance measurements on it and um yeah it was ugly. I mean it's
35:54 just it's hard uh it's a hard problem um especially for larger stacks
36:02 libraries um so uh yeah so I came out
36:11 um beta 10 was a few versions after uh last year's conference um but it was
36:18 still a public version and I added in there that it would cache some of the
36:25 information about your stacks library um into a separate uh yeah data object
36:35 that it could store in a nice place and then it could load that up very quickly so you could see
36:43 um you could see your stacks library immediately. Um, unfortunately it didn't
36:50 actually work deeply because a lot of stacks are are encrypted. Um, a lot of
36:57 them have lots of templates and um before you can actually use things, templates need to be compiled. Uh, and data needs to be decrypted.
37:11 Um, those things take time. So, uh, you know, the pro is that it felt immediate.
37:17 Like it just showed you your stacks library. Bam. Um, the downside is the caching information was just incomplete.
37:26 And so, even though you saw the stacks, if you tried to use them, it just wasn't going to work until the stacks library
37:33 loaded. Furthermore, I had like gotten rid of the loading uh user interface, and
37:40 so I was just very confusing. Um, it was not a good version. Um,
37:49 so I'm going to fast forward uh there were a lot of private versions between
37:54 10 and 20, but um, yeah, fast forward I
38:09 um, it does the loading actually in the background. It shows you edit mode
38:16 immediately. And it does this by knowing what stacks are going to be displayed in edit mode and loading those first.
38:28 Um, the bonus was that you could start working immediately. Um, the downside is
38:36 that there were still stacks loading in um, over time and doing it this way,
38:43 giving priority to the edit mode and loading things in the background actually slowed down the loading. So,
38:51 um, in my demo here, it went from about 13 seconds to load a huge library to
38:59 about 20 seconds. So, I mean, it didn't just get a little bit slower, it got a lot slower. Um, now most of the time you
39:07 you're editing the the stacks that are already on your page.
39:12 Um, that worked pretty well, but if you needed to interact with other stacks, they were still slowly chunking in. Um,
39:20 so that was uh not great. Um, but I, you know, doing it in the background seemed like the
39:28 right thing to do. So I've I've pushed forward through through this. Um, so beta 30. This is maybe a month ago.
39:41 Um, it made one change to this uh, which was to begin loading on launch. So if
39:49 you opened up stacks and
39:54 [Music]
39:55 um and didn't have a project open, it was still loading in the background. And so sometimes if you're fumbling around
40:05 finding where the file that you're going to open, um, by the time you you get things open, if your stack library isn't
40:14 gargantuan, it'll already be done. And, you know, you can just get to work, which is that's pretty nice, super
40:22 fast. But there's a downside. And the downside is that if you happen
40:29 to immediately open a project, then it loads your stacks first. If
40:37 you load the whole library and then open your project, then it's already done and you
40:43 get to work. But if you do anything in between, and it just so happened one of my beta testers was doing things in
40:52 between that way basically every time, it meant that the stacks that were in front of you on the page were just like grayed
41:07 when the library started to load.
41:11 So back to the drawing board, right? Um, so yeah, over the past, I don't know,
41:20 two weeks, if you've been following me on Masttodon, um, I've been posting about this because,
41:26 uh, yeah, I discovered two huge huge things. Good enough that I threw
41:35 away my last item on on this demo so that I could show you this this stuff in detail.
41:42 Um and uh one is that
41:47 [Music]
41:48 um the new version of the security library that came along
41:57 with my FTP library um allows for parallel decryption which
42:07 was not an available thing in the past.
42:12 So suddenly um we went from being able to do one decryption and
42:19 load at a time. It depended on how many stacks you had in your library that were encrypted. But you know if you have a lot of Joe Workman stacks, they're
42:27 pretty much all encrypted. So um yeah, that slowed down. It was like a really huge bottleneck opening that wide
42:35 open. Um it just changes the game fundamentally. But then secondly, I I found another uh detail.
42:43 Um it was very frustrating because I I was loading um I was running testing
42:50 things in Xcode and I realized that my
42:57 Xcode builds were faster than my publish builds, which is a real head scratcher.
43:08 things working inside of a debugger inside of Xcode are fundamentally slower. And that's not
43:16 even when I'm Yeah, they're just slower. So having it be substantially faster, faster
43:23 enough that you could notice it um was a head scratcher. And I dug down and
43:30 did yeah a while of um performance tuning on it and I found
43:39 eventually found the one little quirk where in debug mode only um I was
43:49 verifying my security key once in the release builds I verify it on
43:57 every stack, which is useless. There's no reason. It's the same key. I was doing it every 200 milliseconds. Uh I
44:07 think it would stay the same. Um there's not any security risk there anyway.
44:14 Um yeah, so I added these two things.
44:17 One, it I mean running on 10 threads, you don't really get 10 times the performance, but you get something pretty close. Um, and then yeah, I
44:27 bought another 50% by not running that security verification over and over and over
44:34 again. By the way, that's not security verification on the encryption or on the validity of the stack. It's just running
44:44 uh checks on my own security key.
44:50 Um, yeah, which is good. you should run it at least once, but um doing it over and over again in a short period of time is just pointless.
45:01 So it it doesn't hurt anything. Um and it buys us a huge amount. So I you know
45:08 uh we reduce the number of things that you have to do before you get to work.
45:12 So that's uh by doing things in the background. Then we got like
45:19 a let's say between five and tfold uh performance boost and then another 50%
45:26 on top of that. And these things are multiplicative. So we went from loading in 13 seconds to 20 seconds and then
45:35 back to like well you'll see um it's just crazy
45:41 fast. It's it it's embarrassingly fast because a month ago it really wasn't.
45:49 And uh it's amazing that two very basic uh well one is a new
45:58 feature a new library new version um and another is is this bug that I found um by doing those two things we really get
46:07 this yeah vast change. So, um, yeah, here are the times that I just recorded and I'm going to show you these demos.
46:18 Um, so one, uh, we started off back in the stacks plugin. It is about 13 seconds to load, um, Joe's library. So,
46:28 um, I actually have a library that's about twice this big. Um, it's over 4,000 stacks. Um, and these are
46:36 individual stacks. So, uh, like when you have foundation, it's like 100 or something stacks. Joe can tell you what
46:44 the number is, but, um, it's a lot of stacks. Um, and I'm, yeah, I'm saying how many individual items there are. So,
46:54 uh, for this library, 2631 stacks, 13 seconds in the
47:02 plug-in. um beta 10 from last year. It was really hard to tell because it loads
47:09 and shows you the the stacks immediately, but I know that they're not actually useful. If you try to actually use them, they show up as
47:18 like, yeah, this stack is not installed because they haven't loaded yet. So, I just put question marks here. But, um
47:26 the reality is that it's probably pretty close to the 13 seconds number, maybe a hair faster.
47:34 Um, then we jump up to beta 24 and that's when we're doing it in the background. And
47:42 um, I've added a couple more things to make it a little bit faster, trying to
47:49 squeeze the the blood from the stone a little bit more. Um, and so, uh,
47:57 yeah, I got it better to 20 seconds. Um, unfortunately, if you load your project
48:05 file after it starts loading the library, edit mode doesn't even become functional for another 20 seconds.
48:13 So, I thought I had been doing it was again you make measurements, you make some assumptions about how it how it's
48:21 going to work. Um, and in the end when you actually use it on somebody else's project and somebody
48:30 else's library, not as great.
48:37 Um, version 33, um, I got the edit mode, um, a little
48:44 bit better. Uh, and I was loading that in advance, so we got it down to like five or 10 seconds.
48:53 um before uh edit mode became usable, even if you were loading your project file after.
49:09 about I noticed the quirk about running in Xcode being faster. And along the
49:16 way, I did some reading and testing on how many threads I could run decryption on. So, uh, both of those things
49:26 happened between version 33 and version 35 that I'm showing you today.
49:32 Um, beta 33 was only like last week. So, this is really brand new. This is the cutting edge, but it's so great that I
49:41 have to show you. Um, so version 35, how fast is
49:47 it? Edit mode is fundamentally instantaneous even on a library that's
49:55 2600, even on ancient hardware. So, uh, Joe has a
50:03 2016 Intel-based laptop. They were kind of dogs even when they came out.
50:09 Um, frankly, uh, Apple's hardware team has really been kicking butt since the M M1 and,
50:18 uh, machines are so much faster now. So, um, even on that Intel machine, we could only ever see the
50:26 progress bar flash for a moment before it disappeared. Um, 100% loaded. Well, on this machine, it's about two seconds.
50:35 Um, and that's worst case scenario. So, um, without further ado, let's go check it out.
50:43 Um, right. I have to get rid of this thing that's in my way. Can I put it over there? Yes. That was a zoom toolbar. I know you guys can't see that.
50:55 Um, all right. So, I'm going to start with the oldest one first. Um, yeah, I don't normally uh demo the plugin much anymore these days.
51:09 Uh, kind of left those guys behind. Uh, I mean, not my customers, but
51:16 um, I don't do much Rapid Weaver, although I do have did a beta like two weeks ago, so I guess I am. Um, yeah, let's do it. Here we go.
51:29 Uh, I want this version. I like this version. All right. So, here's Joe's
51:37 library. Look at it go. Wow. It's
51:44 slow. And I know everyone who's here probably knows all about that. So, I
51:53 just want to say I apologize. Uh, it's going to get better. Um, I
52:01 I there are reasons that I couldn't have done the things that I'm going to show you today. The largest part being
52:08 the the uh multi-threaded decryption, which is just not possible in the
52:14 plug-in, unfortunately. But, um, yeah. Uh, well, that's enough of that. We don't need to look at that app
52:22 anymore. Um, and I have somewhere. What is that? No, go away.
52:31 Don't want to see Xcode. Um, I want that. Okay. So, here
52:41 we have a few versions. And I'm going to start with uh beta 10. That's last year's beta. And I'm going to drag this
52:49 file on there so it opens. And that shows you the worst case scenario, which is, hey, look, all the stacks are loaded. They're really not. That's the problem.
52:59 Um, and if you, yeah, have a giant library and you're using a lot of those
53:06 stacks that are loading later in that library load, um, you're going to see this sort of thing and then you click on
53:13 it and it goes, "Oh, wait. I maybe I do have that." And everything that you touch, it just gets weirder. Um, if you
53:21 move something around, uh, like if we move this up here, it's going to Whoa, I
53:28 have everything. Amazing. Um, so yeah, it was a failed experiment. Um,
53:37 I thought pre-caching the stuff would be a great win, but there's only so much that you can
53:45 cache. Um, if I cache all of the library, then really it's just slow to
53:52 load again. Um, so that's when I decided to kind of try
53:59 to prioritize things a little bit. So, um, yeah, let's see how that looks.
54:04 Close that. And this time we're going to load this guy into this version. All right.
54:13 So, uh, you can see down here. Oops. Uh, turn that off. Doc, go away. There. Um,
54:21 you can see it has a library loading and it's happening in the background. And if you watch here, it's populating your library. It's actually changing on the
54:29 fly, but because it's doing all that work in the background, it actually slows it down a little bit. Um
54:37 cuz yeah edit mode and the interactive stuff on your window that will always get priority over other threads. I mean
54:47 you can you can invert that but that's worse. You want to give priority to the the things that are interactive but um
54:55 if you noticed if I did it that way um my page pretty much loads immediately
55:03 but it's variable. Um if you Depending on how you use it and how you open the file and where it is in the
55:11 loading process, it's going to be either usable or not usable or Yeah. Let's Let's see if I can make it
55:21 happen. This No, it's going to work again.
55:26 Um, yeah, I think it's better actually in the next next one. So, let's close that. Um, yeah, I'll just open this one
55:35 like this. So, you can see this one.
55:40 Um, yeah, it's not going to load these stacks until the library is finished loading.
55:49 So, it's behaving a lot like the other one. It's a It's kind of in between.
55:55 Um, so you could still interact with it, but you couldn't actually like set change the settings on the stacks until
56:04 the whole library loaded. And um, yeah, there I could have rewritten that um, and that was my plan, but then
56:13 I discovered these other uh, benefits, these other um, improvements and yeah, I
56:20 decided to go with that instead. So, um, yeah, let's let's just, um, I'm
56:29 going to drop this right on here. See how it goes. Um, it's done. That's it. Whole library is loaded. Yeah.
56:43 Uh, again, I apologize for the status quo because I wish I could have brought that to you guys 5 years ago. Um, of
56:52 course I couldn't. I I needed my own app to do most of it, but um, yeah, it's it's pretty phenomenal
57:02 change. Um, and it's not like I'm faking it. It's that's that's not a cache.
57:07 That's not loading some first and some second. It is actually doing that, but um doesn't need to really.
57:16 Uh it's actually loading the entire library there. And my machine that I'm showing this to you on is an M1 MacBook
57:25 Air. So, it's a 4year-old, the lowest performance in the line.
57:32 Um, I did an M1, uh, which made it a lot faster than the Intel machines, but actually my last Intel machine, which I barely use anymore.
57:42 Um, is yeah, about 50% faster than my M1 MacBook Air. Um, so this is what I'm
57:50 saying is this isn't the fastest machine. If you've bought a Mac in the past 2 or 3 years, it's going to be even
57:57 faster than this. Um, and more than likely you don't have Joe's gargantuan
58:03 library unless you're Joe or Bill or Matteas or there's like a handful of
58:10 guys that have like every stack. And I mean, God bless you. Thank you. uh you
58:18 you make the the platform what it is, but um there's only a handful of people that have that many stacks um and put
58:27 them all in their library. Um cuz yeah, there's some old janky stacks in there that you might not use anymore.
58:36 Um anyway, what are we at? Uh 620, I think I ought to call it.
58:41 Um, it's uh that that's enough of that. Um,
58:48 so yeah, let me just do that one more time. I want to do that one more time cuz it's that good. Come on.
58:57 Look. Boom. It's done. That's it. It's like two seconds. You can hardly even see like one update of the progress bar.
59:07 It's It's insane. It's insane.
59:10 um on my own personal stacks library that only has about a 100 or 200 stacks in it. I mean above and beyond the 40 that are in the built-in set.
59:24 Um you never see a progress bar. It just like it goes loading done. It doesn't update at all. It
59:32 doesn't have time to. Um, yeah, it waits 200 milliseconds before it does the first update of the progress bar and the
59:40 progress bar takes another 10 milliseconds before it actually draws to the screen. So, a lot of
59:47 times in that 220 milliseconds, um, there's already another thread
59:55 coming along and saying, "Yeah, hide the progress bar. You're done. We're done here." And believe it or not, when it
60:03 does that, it's done even more steps in the progress than any of the other versions
60:12 that I've shown you. So, I'm not faking it in any way. This is the full load. No
60:18 caching, the real deal. So, uh I think you're going to like it.
60:24 That's all I'm saying. All right. Um I call it quits there. Um, I'm doing a talk tomorrow on themes. Um, that one's
60:33 going to be really technical. So, uh, bring your thinking cap and we're going to build a theme together or we're going to convert a
60:41 theme. I'm not going to build one from scratch. That would be that would be kind of insane. But, um, I mean, you could do a simple one, I guess. But
60:49 uh I Joe has lent me um his Reason Pro
60:56 uh theme from the other guys uh platform.
61:02 And um yeah, I want to show how
61:08 uh how the theme API is designed specifically to convert themes from lots
61:17 of different platforms. Rapid Weaver, yes, sure. Um but also things like Bootstrap or WordPress. Um not so much WordPress.
61:29 That's a their theming is pretty quirky, but um yeah, a lot of the newer themes for React and whatnot.
61:37 Um Vue and all those those new fangled uh web development platforms. A lot of
61:44 them have pretty simple templating and um we designed the theme API to be highly
61:52 flexible so that you can often times pull in a bunch of those templates, not change
62:09 on what I showed you today or anything really, um let me know. Joe, sweet. I I
62:16 I answered a bunch of questions live. Uh there was a couple there, you know, some simple ones that I answered like, you know, can we export locally? So, if we
62:23 want to FTP via transmit, obviously you have there's an export button on the toolbar. uh the toolbar. There's a command key. You can do the old uh command P and uh it'll export here.
62:35 Yeah, I have garbage stacks on the screen. Sorry. Um uh it's not a great site, but um yeah, export locally.
62:44 There's a built-in PHP server um that yeah, can do PHP and all that
62:51 fancy jazz locally. Um you can preview in the browser, you can preview in the app. Um, and you can publish to a site.
62:59 So, uh, and export to your disk.
63:04 Yes. Sweet. We're only supporting one export location and one FTP location in this first build. Um, that will
63:12 obviously change. Uh, we're trying to get something functional before we build up, but um, that's an easy change. It's
63:22 just faster to get to the end. Uh, if if I only do one for this time. Anything else? People saying,
63:31 "Sweet. Now they can have 10,000 stacks installed and still not have worried about really." I I mean, I think
63:39 um Matteas sent me his uh stack library a few years ago and I think I think he
63:46 has the biggest in the community. Yeah, exactly. Um and I measured it earlier today. Um, it might still be on my screen.
64:07 that, but you know, that's about twice as big. Um, and I suspect it'll still load in five or 10 seconds. I I don't think it'll be a big deal.
64:18 Yep. If you want to watch. Pretty amazing. I can do it.
64:27 Score. Well, that was it. Just a lot of people loving the uh uh the performance and all the other cool stuff. Um people
64:34 can't wait to get their hands on it. So, yeah, I'll get it build out as soon as I can. Um
64:42 yeah, there's a couple things. Uh, I've mangled a lot of the theme stuff for
64:51 um the theme demo I'm going to do tomorrow. Um, so like you notice, I was just using the blank theme in this demo.
64:58 And I haven't even touched preview other than that preview. Uh, oh, you guys can't see that cuz that's on the other
65:06 screen. Um anyway, uh yeah, my my my themes are a little
65:13 bit messed up and my uh that link um the link box has two
65:22 problems. One, a bad crasher um that's easy to bump into. And two,
65:28 um yeah, it wasn't indexing other pages. Oh. Oh. Oh, it
65:36 wasn't um if you like typed
65:41 in a link to an off-site page and you just hit return, it would
65:49 just like cancel. Um, so vaporize your links. Um, which I thought is just that's crappy and it's a trivial fix.
66:07 yeah, you whenever you do a demo, you discover so many bugs. It's just the nature of the beast. Anything else?
66:19 Um, one is, uh, are we going to see from now on kind of more public open public betas?
66:27 Um, what's the plan there? That that question came through anonymously.
66:32 [Laughter]
66:36 It's from Joe. It was not
66:46 Joe. Yes, more, but that's not a high bar to cross when you haven't been doing
66:52 any demos. Um, so, uh, definitely there will be more for theme developers. Um, that's been my biggest hold up.
67:05 Um, that I've been working on the theme API and as soon as you release an API, it's written in stone, you know,
67:14 um, because I if somebody builds something on top of that, I have to support that forever more. And so there's a few things uh especially in
67:22 the layouts themes that I I wanted to get right before I started publishing
67:29 regular updates. Um Mhm. there are other reasons too but I don't want to talk about negativity. Do you have
67:37 uh obviously you can't give an exact okay but like do you have an idea of a launch date? That's it was a question it was asked. Yeah, I gotta ask it. No.
67:49 Um, every time I try to give a schedule, I disappoint 10 people at least and more
67:57 than that if I do it in a public forum like this. So, um, I'm going to say no.
68:02 No, I don't have an idea. I do have an idea, but um, I'm just not going to be public with it.
68:09 Uh, I do feel like we're getting closer to the end. Um, but there's still work to do. Yep.
68:15 people. So, you said potentially a beta hopefully tomorrow. You want Yeah, this weekend. I mean, obviously I'm giving a
68:22 talk tomorrow, too. And uh it's lots more technical and I'd like
68:29 a few hours to practice and I did most of the demo work for that in the past
68:37 week and yeah, I haven't practiced a lot. So, we'll be winging it a little bit tomorrow.
68:46 I I I'll give a little PSA to everybody.
68:49 You know, once you get this, feel free, you know, we don't want to bombard Isaiah. Isaiah probably is going to rebuke this and say, "No, everyone come
68:57 to me cuz, you know, he's going to feel responsible to help everybody out." Hey, I'll take the brunt of the hit. You come to me and I will gladly help you out as
69:07 much as I can. Um, and uh, you know, be the front man. so uh our our godfather Isaiah here can uh finish this and get
69:15 it uh get it out to the world as soon as possible. Um I do already have a STAX beta space set up on my community. Um so
69:24 for those that have access, it is open to the public. It's not um so if you want to join that STAX beta and you have it, you can go there maybe, you know,
69:32 talk it up there. Um and I'll be glad to help you out with anything that I am aware of. Yeah. just not aware of
69:41 everything obviously. Uh yes, but uh he does bombard me with questions all the time
69:48 and uh it's all love. It's love. I promise. Yeah. Yeah. Uh anyway, um he knows a lot of details um and has Yeah.
69:59 obviously seeing more bit of more I mean there was a gap in there of
70:07 quite substantial and Joe was very patient not okay no he wasn't very patient but he was patient a little
70:15 patient kind of patient I I had my patient days you okay but he you know he keeps me going
70:22 you know every day encouragement uh and uh even if I don't always respond
70:29 to his his texts. Um it's it's good to have that encouragement. Um even and Amazon doesn't let me buy ibuprofen for him anymore.
70:42 Yeah. Uh I Yeah. I I you know, I wanted a a moment to mention about my health cuz it was a big deal. Um and last year
70:51 I gave a whole talk about um where I was in recovery. Um, and to
70:58 be honest, it's been 18 months now since I hurt myself um on a skateboard.
71:04 Um, and I can see now in retrospect that I
71:11 lost a lot of time. I mean, there were months that I didn't do much of anything. And even by last year's demo,
71:20 I had only been working like full-time days and just barely full-time days for a couple of months. Um everything before
71:27 that I was doing you know voice to text and boy that no way to program.
71:36 Yeah. You can't get much done that way. So I I mean I worked on other things. You're no Stephen Hawking, you know.
71:44 Well, I mean I could do it like that. It just I actually saw Stephven Hawking talk once. Um, I was fortunate enough my
71:54 dad knew that he was going to be a thing and took me when I was very young. Um, and he took questions and answers at the
72:02 end, but it took like 2 hours. And every question, somebody would ask it and then he would pick one to answer.
72:12 and you know, he'd type it out with his basically could move his neck a little bit and that's how he uh talked.
72:21 Um, well, I stayed for the whole thing.
72:28 Very cool. Very cool. Anyway, uh, thanks for listening. Thanks for, uh, being patient with me. I know not everybody's
72:35 patient, but, um, for those of you who are, it's really appreciated. Uh, I am working on this, you know, 10 to 15
72:44 hours a day, every day. I don't really take days off. Um, and I've been doing it for, you know, few years now. So, um,
72:52 we'll get there. I don't think we're that far anymore. I hope we're not that far anymore cuz I'm tired. You know, I I was thinking recently it's, you know, I
73:01 know it's easy to think of, you know, stacks as this app as being like a 1.0 app. let's just get at what we need out.
73:09 But we can't do that, right? Because that's the big because this is it. We we need to go from zero to a a mature
73:16 15-year-old app, you know, because that's what we need to to continue going forward. You know what I mean? And yeah,
73:24 and that's not a simple thing that has turned out. I mean, it's the been the biggest discovery is realizing
73:34 that the app can't just be like a superficial thing. It has to be deep and
73:42 support compatibility with the full stacks API, which means uh recreating some things that stacks
73:51 the plug-in leaned on rapid for for years. Um, and although I became less dependent on those things over time,
73:59 there's still some pretty significant things. Um, and then developing a theming API on top of that, I I didn't
74:07 really realize honestly at the beginning that it would be first kind of a challenge. Um, but
74:17 then also along the way I would discover, oh, like somebody's going to want to convert a theme that looks like
74:23 this from handlebars or from a mustache template or whatever. Jackal or some Jackal. Yeah, that's another I use
74:31 Jackekal as as a demo conversion all the time. Um, so, you know, I had to support a lot of flexibility in the options.
74:39 Yeah, that added a bunch more complexity. And you know, every time I poke at something, it feels like, oh, I can't just do the superficial thing.
74:50 Yeah. Yeah. Like, uh, FTP is a great example where I could do the superficial thing. Um, you know, I just like I'll just do one location.
75:09 and I already had an app that did that, so I just stole it from that old app.
75:17 Um, like, uh, Joe mentioned in the previous talk, he has a process for building his app and notorizing it and deploying it.
75:28 And I have that, too. Um, I have a little app called deploy. So, yeah, I stole that. and the idea for calling it deploy anyway.
75:39 Um yeah, sometimes I I could do the superficial thing sometimes, but um you know, other times it was
75:48 like like the PHP server that's built in there. Uh you know, I had to build it and then make sure it supported Intel
75:55 and and uh M1. And Joe just gave me a bug report that says it needs to be recompiled with these five other plugins
76:04 in it. Um, so again, you know, doing the superficial thing would have been fine for a brand new app, but because people
76:12 have all these expectations of what they used to do and what they needed to do for their stack or for their project or for their company.
76:23 Um, I kind of feel obliged to do the full thing or at least uh get close to the hole.
76:31 Yep. Sweet. Well, bud, thank you very much. I appreciate it. I appreciate you and uh definitely looking forward to tomorrow. When is the port to Linux?
76:41 Soon. I I um four years ago. Um
76:51 when I got the first builds of
76:56 the new Rapid Reaver API that just sent us into a tail spin. Um, I got very fed
77:05 up and I learned NodeJS and Typescript and I built a full version of the STAX
77:14 API with rendering in Typescript. So, I mean theoretically I could start there but you
77:21 know just like this one building the full app would take a lot. So, um, Linux is not soon,
77:31 but I'm thinking about it, honestly. Um, I use Linux a lot nowadays, a lot more than I used to. Ah, I see iPad OS first.
77:39 I I know you love iPad OS there, Isaiah.
77:43 [Laughter]
77:46 Not happening. Um, not unless something really fundamental changes. Uh, iPad and, uh, iPhone means going through the app store. That means giving Apple 30%.
77:58 That means a whole bunch of other limitations. Um, by staying off the app
78:04 store, it gives us so much more freedom um to do a bunch uh do a bunch of other
78:11 things. So, um maybe someday I'll go to the app store. Maybe the app store will change. Maybe uh in Europe I could do
78:19 one of the like side app stores, you know, those have different rules. Who knows? But um yeah, I'm not
78:27 thinking about app store really for now. VS Code Remote. Yeah, maybe an Electron version.
78:39 Fundamentally, it could be ported to How's
78:47 it? Um fundamentally it could be ported to a web app, too. Um, but that has other other parts of it because
78:55 um, yeah, I mean VS Code uh, add-ons are usually open source. Um,
79:04 there are some that charge subscription fees, but uh, yeah, the business model is very different for those.
79:12 Um, if I built for a web app, then I would definitely have to charge some sort of subscription fee because you're,
79:20 you know, running a service at that point. Y um, that's a whole different other beast. So, that's probably not in
79:28 the cards too soon, but this it's not out of the question.
79:35 Cool. All good. Trying to read them as they flash by. Yes, exactly. Yeah. Yeah.
79:42 Yeah. Cool. Well, uh we can continue this on the hangout afterwards. 40 extra minutes. Um Yep.
79:51 Oh my gosh, there's a lot of comments.
79:53 Um yeah, thanks Joe. Thanks for having me again. My pleasure. I'm about to pee myself.
80:04 That's the kind of response that that deserved. Damn it. Thank you. Thank you for getting
80:13 it. That That's exactly what you were supposed to supposed to do.
80:20 Nice. Cool, dude. Take care. I appreciate it. Take care. Bye.
80:30 [Music]