0:00

Welcome to syntax on

0:02

this Friday separate club. We have

0:04

Carson gross here. We're going to

0:06

be talking all about HTM X.

0:10

We're going to be able to pick the

0:12

brain of the hyper media man

0:14

himself and talk all about everything,

0:17

uh, that you all have

0:19

been asking about since our first HTM X

0:21

episode, my name is Scott to lens game,

0:23

a developer from Denver with me, as always

0:25

is West boss. What's up, Wes? Hey,

0:28

uh, not too much. Excited to have Carson on

0:30

today to ask him all the questions

0:32

that we have about HTM X and, uh, understand

0:37

what's going on here and hopefully clarify. Cause it's

0:40

like, we had tons and tons of feedback from

0:42

that show. And some of the

0:44

stuff was valid and some of the stuff was just like,

0:46

oh, you actually, I don't think that you totally understand how

0:48

it works. So hopefully we're going to clarify all of that.

0:51

Yeah, totally. And, uh, also

0:53

if you're asking questions of your code, like,

0:55

Hey, why did this bug happen? Maybe you

0:57

want to have a tool like century on

0:59

your side to reveal all of those bugs

1:01

and why they happen so that way you

1:04

can go solve them. So if you want

1:06

to use a tool like that, I had

1:08

on over two century.io/syntax sign up

1:10

and you'll get two months

1:12

for free. Carson,

1:14

how's it going? My man. Uh,

1:17

it's going pretty well. Uh, a little

1:19

crazy things have been crazy since, uh,

1:21

last July in HTM X land for

1:24

sure. Um, got a lot more attention than

1:26

I was certainly used to. Um,

1:29

when the

1:31

primary gen picked it up and fire

1:33

ship dev, um, did a short on

1:35

it. And so there's been a sort

1:37

of a huge amount of attention, um,

1:39

little overwhelming, but you know, it's

1:41

all good. It's fun. Yeah. Was

1:43

that the sort of catalyst for the, it

1:46

was, I was sitting

1:48

on a, on a bucket at

1:50

a, at the state youth baseball

1:52

championship. Um,

1:54

uh, baseball coach here and, uh, I just

1:56

checked Twitter and was like, that can't be

1:59

right. And

2:01

it's kind of it's been that

2:03

can't be writing for since about July. It

2:05

was really good I got very lucky because

2:07

I released a book like you know I'm

2:11

at the end of July middle and end of July last year

2:14

And so it really helped just a lot of

2:16

things kind of came together at one time So

2:18

very lucky in that in that sense Cool.

2:22

Well, do you want to give a brief introduction

2:24

to who you are and what you do

2:26

to our audience? Yeah,

2:29

I so my name is Carson gross I

2:31

run big sky software, which is a

2:34

big sky dot software and That's

2:37

just a consulting company that I run that supports

2:39

my open source work for the most part And

2:42

then I teach my primary job is I teach

2:44

at Montana State Which is where

2:46

the primogen actually went for under that nice

2:49

So funny connection. Did you teach him?

2:52

I didn't I you know, I did not I

2:54

know people who taught him though There are people

2:56

that are still there that he that he knows

2:59

So all of these hdmx ideas

3:01

are like institutionalized at a very

3:03

high level. Yeah, we're trying

3:05

I mean, I'm yeah, it's I'm trying

3:07

to write actual academic papers on it

3:10

and we'll see Academies

3:12

kind of its own things. So So

3:15

you're out in Bozeman. Yeah, Bozeman, Montana.

3:17

Yeah, I've been there a few times.

3:20

Yeah Cool town the

3:22

joke is that the nice thing

3:24

about Bozeman is like it's right next to

3:26

Montana You

3:29

can see you can see Montana from Bozeman so

3:33

There's some just I don't know that the

3:35

vibe out there is really great Obviously

3:37

that the skiing and mountain stuff is wonderful

3:39

too. But yeah, big. Yeah Yeah,

3:42

it's cool town. Don't move here, but it's

3:44

cool town Yeah,

3:47

I get it And

3:49

like what led you to Developing

3:52

what hdmx is actually maybe let's step back one

3:54

step further because people are probably tuning into the

3:56

first couple of minutes I mean like should I

3:58

listen to the rest of this one? Read.

4:00

What is Ht Max for anyone who

4:03

hadn't heard of it? Sorry, she makes

4:05

it So basically you can think of

4:07

it as Html Plus plus it's it.

4:09

Takes the ideas of Html and particular

4:11

anchor tags and forms and I'm what

4:14

I say those are. to use some

4:16

fancy terminology those are what are called

4:18

hyper media controls. One of the interesting

4:20

things about Html as the you embed

4:22

these controls the things the things we

4:25

do stuff inside the document itself, arm

4:27

and a sorry what is too much

4:29

does it takes us. To ideas,

4:31

forms and legs and then tries

4:33

to generalize that idea. So I'd

4:36

is it lets you put attributes

4:38

on it in Html that led

4:41

you trigger are an http requests

4:43

based on whatever of that you'd

4:45

like and then that issues be

4:48

requests is going to be returning

4:50

Html rather than juri sandwiches was

4:53

normally. Used for

4:55

for Ajax request used as and

4:57

then at a similar was placed

4:59

in that dom somehow. Com

5:01

has reached efficient as basically just

5:03

provides you. There's like seven or

5:05

eight core attributes that let you

5:07

generalize that idea and others The

5:10

surprisingly powerful seven or eight attributes.

5:12

He knew quite a bit with

5:14

it. So if list serves, want

5:16

to go raised your max.org/examples we

5:18

have like infinite Scroll can be

5:20

implemented in those terms active search

5:22

were like two types I'm playing

5:24

and then the results kind of

5:26

immediately come back There can be

5:29

implemented. he's three or four. html

5:31

attributes and it's all done in

5:33

terms of html rather than sort

5:35

of or bridge and for at

5:37

a thicker client side model on

5:40

such that's the big idea their

5:42

harm it's again i think it's

5:44

probably submitted sounds pretty simple in

5:46

it isn't super complicated temple many

5:49

other arm but it's a surprisingly

5:51

powerful idea i think that hyper

5:53

media is under us made by

5:55

most people today because html has

5:58

unfortunately to sort of stopped making

6:00

progress in that area. And

6:03

so people when they think about hypermedia,

6:05

they think about clunky full

6:07

page refreshes and so forth. All

6:10

the browsers have gotten better at that stuff. One

6:14

thing that I really noticed is that when

6:16

you bring up HTML people are so quick

6:18

to say, well, you can't do X, Y,

6:20

and Z with it. And I

6:22

do think that one of the

6:25

smartest things that you've done on the site

6:27

is have this examples page because if people

6:29

were to just take a quick glance at

6:31

this examples page, you can

6:33

see that they're really, that kind

6:35

of invalidates many of those conversations because

6:38

people just straight up, they

6:40

underestimate it like you said, and they underestimate

6:42

what it's capable of, especially in

6:45

the fact that many of the times when

6:47

we're doing a lot of this stuff, deleting

6:49

a row or like you said, searching, you're

6:51

having to hit the database anyways, right? You're

6:53

having to make a server trip. So what

6:57

are we doing just sending JSON and then using

7:00

JavaScript to distribute that? And

7:03

so I think that's really powerful.

7:05

I wanted to know like what

7:07

was the catalyst that made you

7:09

want to create something like HTML? Yeah,

7:13

so the original HTML was intercooler

7:15

JS, which was released back in

7:17

2013. And

7:20

that was based on some work that I had

7:22

done internally at a startup for a friend where,

7:25

and it's kind of funny, but I

7:28

was trying to sort a table and I was

7:30

trying to do it client side because in 2013,

7:32

all the cool kids were doing stuff client side

7:35

and JavaScript. I was using

7:37

jQuery because in 2013, everyone was using

7:39

jQuery. And I

7:41

just couldn't make it fast. I was trying

7:43

to sort a table, there were a couple thousand

7:45

rows in it and browsers at the time, I'm

7:48

sure now browsers can handle it without breaking a

7:50

sweat. But at the time, the

7:52

browser runtimes were not as good. And

7:54

so it was very slow. And

7:57

I ended up out of desperation just trying.

8:00

this technique out where jQuery

8:02

had a few methods for issuing

8:04

a request and then slamming

8:06

the results into the DOM. And

8:09

I did that and it was very

8:11

fast. This was a

8:13

Rails app at the time and I

8:15

was a little bit shocked by that

8:17

and so I started writing, I recognized

8:19

that technique. There's also something

8:22

called Pjax which is sort of an older

8:24

take on this idea. So

8:26

there were some ideas floating around and I then

8:29

saw Angular 1 and the

8:31

way they used attributes and I thought

8:33

to myself, well, I'll just use

8:35

this attribute idea, this custom attribute

8:38

idea to basically drive this little

8:40

custom function that I've been using

8:42

to do this same pattern. And

8:46

that eventually grew into intercooler JS. So it

8:48

came out of a performance, what at the

8:50

end of the day was a performance issue.

8:52

I wasn't super passionate about hypermedia at the

8:54

time. It was only after

8:56

I released it that people pointed out to

8:59

me, hey, this is restful, this is like real

9:01

rest. I don't know if we want to get into

9:03

that or not but this is a – the language

9:05

I would use today is

9:07

like this is a hypermedia oriented

9:09

approach to web development that you're

9:12

doing. So eventually I learned

9:14

a bunch more about that and kind of

9:17

wrote a bunch of essays and understood

9:19

what I was doing a lot better over time. And

9:23

the HTMX is basically a rewrite of

9:25

intercooler JS where I pulled out the

9:27

jQuery dependency just because JavaScript

9:30

has made so much progress in the last 10

9:32

years. It was possible to

9:34

do that in a way that worked across most of

9:37

the major browsers. Yeah. I

9:39

remember Pjax first hitting the scene and

9:42

it blowing my mind. Being like, wait a

9:44

second. We're doing a route here

9:46

without a full page load. Wow,

9:48

this is crazy. Yeah. I

9:51

was building like an audio player on a site

9:53

and the fact that I could change routes and

9:55

have that audio not restart without having to pop

9:58

it into a second window was like pretty pretty

10:00

mind blowing. I often wonder

10:02

what HTML would look like

10:04

if you were to totally

10:06

scrap out everything we've done and

10:09

done it from the beginning. Do

10:11

you think that we would have some

10:14

declarative ways to do what we're doing

10:17

to run fetch and get parts back?

10:19

Do you think that HTML might look

10:21

like HTML

10:24

if we were to scrap it

10:26

and start fresh? I don't think

10:28

you even need to scrap it and start fresh. I

10:30

think that the ideas of HTML, I say this a

10:32

lot, like HTML is cool and it's

10:34

not like it's a big library. It'll

10:38

be useful for people that adopt

10:40

it. But I think the idea

10:42

is like the essays, the hgmx.org/essays

10:44

page and then our book at

10:46

hypermedia.systems, which you can read online

10:48

for free. It just talks a

10:50

lot about hypermedia in terms that

10:52

normal developers can understand. Unfortunately, the

10:54

hypermedia world can often be pretty

10:57

academic. There's a lot of

10:59

negativity in it as well, I've noticed.

11:01

I don't know what it's, you know,

11:03

but it's whatever. But the

11:05

ideas there are good and they're interesting. And

11:07

so I think you could take the ideas

11:09

of hgmx and integrate them into HTML in

11:11

maybe a more standardized form. Like, okay,

11:14

maybe you don't want everything to be

11:16

a hypermedia control because of accessibility issues.

11:19

But let's just look at buttons and

11:21

make it so that buttons have the

11:23

ability to be standalone hypermedia controls. I

11:25

think you could start making progress again

11:28

as a hypermedia. There's also

11:30

some like frame sets. There's some ideas that

11:33

hgmx kind of lets you, gives you

11:35

a lot of flexibility around that were

11:37

there in nascent form early on. Like

11:39

frame sets and you can target an

11:41

iframe with a link. It's just that

11:43

iframes kind of stink, they can't participate

11:45

in the DOM. There are frames and

11:47

there are frame sets and those have

11:49

been deprecated to an extent. And so I

11:51

think the ideas are kind of laying around.

11:54

And my hope is that hgmx helps sort

11:56

of spark those, you know,

11:58

spark the thinking around. What could HTML

12:01

be? Because I don't think HTML can,

12:03

you know, Figma, for example, like you're

12:05

not building Figma in HTML, or

12:07

at least you're not building it in, you know, HTML-based

12:10

logic, like there's just too much going on

12:12

there. But on the other

12:14

hand, there's a lot of apps that probably could be

12:16

built in pure HTML, and very

12:19

effectively if they were to start, you know,

12:21

picking off some of the bigger ideas from

12:23

HTML. Yeah, you

12:26

know, it is something that I've been

12:28

really paying attention to when I use

12:30

various apps that I use day-to-day. Like,

12:33

I tweeted about this. It was, you know,

12:35

like how many of these things really

12:37

need anything more than

12:39

HTML to do what they're doing? And

12:42

I don't know if it's that people don't realize that you

12:44

can do things like, you know, streams

12:47

or server-sent events or stuff

12:49

like that, but it's really

12:52

way more capable. But

12:54

the one comment I tend to get on repeat

12:57

whenever I bring this up to people

12:59

is, why would we

13:02

go backwards when

13:05

we went forwards to this,

13:07

you know, client side and now

13:09

SSR and hydration world or even

13:12

resumability world, and

13:14

that it's a step backwards into,

13:16

you know, old-school Ajax?

13:19

Do you have like a comment on

13:21

that kind of response that people

13:24

typically have? Yeah, I think,

13:26

you know, I look at it like

13:28

more of a Hegelian dialect. Like, you

13:30

know, it's a dialectic. Like, we

13:33

were all server-side.

13:36

It's not thin client, but

13:39

hypermedia-oriented, where

13:41

hypermedia was really running the show.

13:44

And then we moved to the

13:47

reactive thick client

13:50

model. And, you know, in some ways,

13:52

I have to say, even reactivity isn't

13:54

new. Like, in the 80s, we were

13:56

building apps in this way, you know?

13:58

Like, we had... you look at like

14:01

Java swing, for example, it

14:03

has reactive aspects to it. There's a lot of

14:05

infrastructure there. And you know, I built I

14:08

built apps like that. Initially, this is kind

14:10

of ironic. But initially, I sort of rejected

14:12

the web. And I built web Java

14:15

based remote app, you know, using what I

14:17

would consider a real remote protocol, as in

14:19

JSON. And so I've, you know, I'm pretty

14:22

familiar with that stuff, or at least that

14:24

was, it's been a long So,

14:27

you know, I think there's there's been more and

14:29

back more back and forth,

14:31

particularly if you step back, then

14:34

probably appears to particularly younger web

14:36

developers. And so, you

14:38

know, I think there's a there's again, there's a

14:40

dialectic here, the original web

14:42

HTML, it was amazing what we accomplished with

14:44

HTML, you know, like the web apps that

14:47

we were able to able to build, it

14:50

just showed the power of, you

14:52

know, the network. And because

14:54

it was clunkier, and a lot of

14:56

people complained about, you know, web apps

14:58

versus the thick apps they were used to,

15:00

but web apps still kind of one and the web

15:02

still kind of one. But then

15:05

eventually, we realized, okay, this user,

15:07

the user experience that we can

15:09

produce with hypermedia isn't good enough.

15:11

So we need to start moving

15:13

towards JavaScript based applications. And

15:16

that's fine. But then there was also,

15:18

you know, in my mind, HTML just

15:20

kind of was always hamstrung

15:23

by not continuing to add

15:25

features, hypermedia features in any event,

15:28

as it made progress. And so I view

15:30

HTML is sort of like a

15:32

not necessarily a return to, you

15:35

know, an older model, so much as

15:38

it's a, it's building, taking some of

15:40

the good ideas from today, you know,

15:43

and then building on some of these older

15:45

ideas as well. And so I think there's

15:47

a dialectic in play there. And so HTML,

15:49

HTML isn't right for everything. There's a lot

15:51

of apps that it doesn't, it just won't

15:53

work for. Yeah. And we have an essay

15:55

up on that when to use when to

15:57

use hypermedia. So, you know,

16:00

I think I just you know, oh, it's

16:02

an endless cycle. I don't know. I mean

16:04

we're making progress HTMX is better than pjax

16:06

in many ways. So we're making progress. Yeah,

16:09

what are some examples of apps that are Like

16:13

you shouldn't build with with htmX. Like when

16:15

should you? Personally, do you

16:17

think that you should reach for react or some

16:19

sort of spa framework? Yeah,

16:21

well, I don't know about react cuz I just to be honest.

16:23

I don't have a ton of experience with it So I don't

16:26

know what it's good at Okay,

16:28

obviously Figma is something you're not

16:31

building in htmX And

16:34

similarly, I you know another example that's

16:36

given in that essay is something like

16:38

Google Sheets would be difficult to implement

16:41

in htmX

16:43

because In Google Sheets

16:45

you've got very fiddly Interactions

16:47

you update one cell in one part of the screen

16:49

and then you've got to update some other cell in

16:51

another part of the screen and that's all highly dependent

16:53

on the Formulas and it's very difficult

16:55

like all you you know with htmX You kind

16:58

of want to refresh the whole thing and that's

17:00

not gonna work in a spreadsheet environment

17:02

now on the other hand I will say if

17:05

you've got like an enterprise application and you

17:07

can do sort of a row editing of

17:10

tabular data Then htmX can

17:12

be very effective like if most of your value

17:15

is being added on the server side Rather

17:17

than on these client-side interactions, which is the

17:19

case in a lot like ERP apps

17:21

and you know any sort of app

17:24

where you're coordinating human or not any

17:26

Coordinating activities like very often

17:28

the the value of the app is is

17:30

on the back end It's not in the

17:32

front end and in those types of applications

17:34

htmX is maybe a better option because not

17:37

because you can do better You UI with it, but

17:39

just because you can do good enough UI and You're

17:43

adding your value on the back end and then

17:45

you save a tremendous amount of complexity Hopefully and

17:47

then that you know that complexity that you save

17:50

on your front end can be dumped into

17:52

Functionality on your back end instead where

17:54

where there's more value to be found

17:57

and do you think that puts more of a

18:00

more of a load onto the server to

18:03

do more. I think that's a comment I

18:05

get a lot. I've been talking a lot

18:07

of HTMX in the Svelte

18:10

discords and stuff like that and people say

18:12

well that's putting a lot of that resources

18:15

now that you need onto the server rather

18:17

than dumping it into the client. Yeah,

18:19

I mean there's some truth

18:21

to that. An HTMX app

18:24

is going to be less

18:27

burdensome on the client because

18:29

first of all we're just parsing HTML which

18:32

is done and we

18:34

don't have a lot of heavy JavaScript logic. That's

18:37

actually typically a good thing because it

18:39

means your client can often perform more

18:41

quickly. One

18:45

thing I've tried to point out to people is that

18:49

the difference between rendering a JSON

18:51

representation and rendering an HTML

18:53

representation particularly if you have

18:55

compression enabled which you should

18:58

is actually very low. There's

19:01

not a massive difference between the

19:03

two. If you have tabular

19:05

data, HTML is actually more efficient

19:08

in some ways than JSON because

19:10

JSON repeats field names unless

19:13

you do something crazy in

19:15

your formatting. I don't think there's

19:17

as much of a CPU and

19:19

memory and network

19:21

transfer time difference between JSON

19:24

and HTML as many people

19:26

suspect. That's

19:29

the first thing that I'd say.

19:31

On the other hand, what HTMX

19:33

does imply is a network request

19:36

for some interactions. Not

19:38

all interactions but for some interactions. There's

19:40

a network latency there

19:42

that is present in HTMX that

19:44

might not be present in a

19:47

well-tuned SPA application

19:50

because a well-tuned SPA application might

19:52

do an optimistic update or whatever

19:55

or might be able to do

19:57

something entirely client-side. That being

19:59

said, there are tools available in HTML,

20:01

like the details element, for example,

20:04

that are there to support client side

20:06

interactivity and htmX is not

20:09

against client side interactivity, per

20:12

se, like we encourage scripting, we

20:14

just say you should script in

20:16

a hypermedia friendly manner, like you

20:18

shouldn't be doing, you basically

20:21

shouldn't synchronize state with the

20:23

server in a non hypermedia

20:25

way. Now that's, you know, I'm pragmatic

20:27

about it, like there's gonna be times

20:29

when you're just gonna do that. And I'm not

20:31

gonna hunt you down about it. It's fine. It's all good.

20:33

We got to get work done. But

20:36

so I think that there, you know, there

20:39

is some truth to that, that there are

20:41

going to be user experiences that are difficult

20:43

to achieve in htmX that are easier to

20:45

achieve in SPA. On the other hand, what

20:48

I would say is, is your experience with

20:50

SPA is universally good. Like

20:53

this, this promise of performance that we've

20:55

gotten, I would say has not been

20:57

achieved as broadly as many of the

20:59

advocates for SPA is would suggest it's,

21:01

it's hard to do, there's a lot of

21:04

layers often between like the front end, and

21:06

the back end, it's, you know, can be

21:08

difficult to figure out exactly what's happening. And

21:11

so that can lead to, you know, the sort

21:13

of famous gif of

21:15

like, you know, an argument in favor of

21:17

SPA is and then it's just a spinning

21:20

gif, right? So, you

21:22

know, SPA is to be done very well,

21:24

I don't want to say they can't. But on

21:27

the other hand, I think you can build very

21:29

good htmX based applications that don't overburden your

21:31

server. htmX is focused on updating partial

21:33

bits of HTML, and partial bits of

21:36

HTML can be very efficient to generate,

21:38

they can actually be more efficient to

21:40

generate than like a bunch of JSON

21:43

API calls to build a particular part

21:45

of your UI. And what

21:47

about your thoughts on

21:49

like a lot of these frameworks right

21:51

now, obviously, react server components, they

21:54

are now saying, all right, well,

21:57

let's render everything on the server, because Because

22:00

in a lot of cases that makes sense.

22:02

Why send the JSON and the library

22:05

to build the table and all this other

22:07

stuff that you need just to render it

22:09

out in the browser, just do

22:11

it on the server where it could be a lot faster

22:13

and then just ship the actual

22:15

endpoint, right? Clearly that's

22:18

going to be, in some cases, in a lot

22:20

of cases, that's going to be a much smaller

22:22

payload that needs to be shipped to the end.

22:24

But you can sort of dip into front-end

22:27

stuff as well if you need it. Do

22:29

you have any thoughts or opinions on

22:32

that trend as well? Well

22:35

I agree. I

22:38

agree. I

22:41

think that those approaches, from what I've seen, and

22:43

I'm not an expert on that stuff, I

22:46

get people that come in on Twitter and are

22:48

like, oh, have an opinion on server components. I

22:51

can't even have an opinion on React really. I

22:57

think as well as I understand it,

22:59

my criticism of that would react components

23:01

in particular. And even this might be

23:03

wrong because there's someone I think who's doing

23:05

work to make this not true. It

23:08

tends to tightly couple your front-end

23:10

and your back-end. Like now you

23:12

have to be running React type

23:14

code or JavaScript on your back-end.

23:16

Whereas if you just take the

23:18

hypermedia model and say, okay, we're going

23:20

to exchange hypermedia. One of the good

23:22

things about HTMX is you can pick

23:24

whatever back-end language you want. That's something

23:27

that I'm pretty passionate about. There's

23:29

been a meme going around, like

23:31

this lava flow kind of

23:34

coming at people and there's

23:37

this sort of fruitless person using

23:39

a backhoe trying to save

23:41

a street that's obviously going to be destroyed. The

23:44

title says this is back-end developers

23:46

trying to prevent JavaScript from taking over the

23:48

world. I think if the

23:51

situation is not that dire, and I've seen a

23:53

lot of people in the Go world, obviously

23:56

the Django world, the Python world,

23:58

pick up HTMX. I think HTML does give

24:01

life to some of these back-end

24:03

frameworks that have been for the last

24:06

five years in particular kind of shunted

24:08

aside in favor of the, what

24:11

a lot of people seems like the obvious pick of

24:13

JavaScript on the server. That's not to say I'm not

24:15

in favor of JavaScript on the server. There's

24:17

some very good run times now. I

24:21

think there's reasons to pick it, but

24:24

I don't like the idea that you have

24:26

to pick it on the server

24:28

side. Even

24:30

that's not a particularly informed

24:33

criticism because again, I've seen

24:35

someone claiming they've been

24:37

implementing React server components in

24:39

other programming languages. It

24:41

makes sense that you can do that. I

24:45

build full-stack JavaScript applications and my

24:47

back-end is 90% not, it's not

24:49

React. My

24:51

logic is not React. I'm not writing

24:53

database queries in JSX. I'm writing it

24:55

in JavaScript and then I pass that

24:57

data to my templating layer.

25:00

That's the point where you sort of pick up

25:03

React. I do think that

25:05

it's important because a lot of times people just

25:07

assume that's the only way to build these types

25:09

of things and they realize, no, there are large

25:12

companies out there that do not

25:14

have a tightly coupled front-end and

25:16

back-end and they need something like

25:18

this. I thought that is pretty

25:20

neat that you can do that as well as

25:22

some people prefer to build their

25:25

logic in PHP or Ruby

25:27

or Python or all these different languages.

25:30

One of the things I say is like, I'm never going

25:32

to be a Lisp person. I may use

25:34

Go at some point in my life, but I'm never going to

25:36

be a Lisp person. I think

25:38

it would be a shame if the Lisp

25:41

community was not able to participate in web

25:43

development because we had just

25:46

decided, okay, we're going to do everything

25:48

in JavaScript. I think it's good for

25:50

people to exist in

25:52

the Lisp programming language to put pressure

25:54

on other programming languages and so forth.

25:57

So I like that diversity of back-end

25:59

choices. for web development,

26:01

which I think HTML helps with.

26:04

Yeah, I think there is some poignant

26:06

things in there. The fact that I

26:09

did hear this criticism from

26:11

React folks a lot when talking

26:13

about HTML was that it does

26:15

tightly couple your server to the

26:17

UI because you're sending HTML. Unironically,

26:22

they're not even recognizing the fact that

26:24

if you are using React server components,

26:26

that's a much more tightly coupled system.

26:30

Given that you're saying that you

26:32

like that it can expand to all sorts

26:34

of different back-end languages, personally,

26:36

like you, if you're writing an

26:38

app, what is your favorite stack

26:41

to work in? Well,

26:43

I worked in Ruby for a long time and there's an

26:45

awful lot of tools in Rails

26:47

in particular that are just there that

26:49

are really nice. I

26:52

don't do, I may start this

26:54

summer doing some web development again, but I don't

26:56

do a lot of day-to-day web development, unfortunately,

26:59

right now, because I'm so

27:01

busy with teaching. But

27:03

I would probably, at this point,

27:06

I would take a hard look at Django. I didn't

27:08

pick Django for the book because

27:10

it's kind of like Rails in that it's

27:12

sort of a one-stop shop that

27:15

has everything that you need, but

27:17

that also adds a bunch of complexity. So in

27:19

the book, I used Flask

27:22

as the web server because it's just a little

27:24

bit easier, I think, for non-experts to pick up.

27:26

And this gets at something that I've been talking

27:29

about on Twitter a little bit lately, which is

27:32

so many of these things are context-dependent. Like

27:34

you shouldn't be a

27:37

false dichotomist or whatever. Like sometimes

27:40

one thing's the right choice and sometimes another

27:42

thing's the right choice, and it just depends a

27:44

lot on the context. So

27:46

I would probably lean that direction. A lot of it

27:48

would be driven by exactly what I was trying to

27:50

achieve. Like if I was doing an AI thing, I

27:54

would definitely look at Python. If I was doing

27:57

something that was maybe... Let's

28:00

see what's another good example. Like

28:02

you, Java's got a ton of really good libraries. I

28:04

know Java pretty well. If there was a good library

28:07

in Java that did the thing that I wanted to

28:09

do, I wouldn't be afraid of picking it and

28:11

using like Javelin or something like that for the front

28:14

end. So I think it

28:16

just depends, just depends. Cool.

28:18

Yeah, I've been personally grabbing all

28:22

kinds of different stuff to work in HCMX lately, whether

28:24

it is like gave me an

28:26

opportunity to get my hands dirty with a

28:28

bun, the JavaScript runtime a little bit, just

28:30

because it was easy to

28:32

pick that up and shoot some

28:35

HTML. It comes with a folder-based

28:37

routing setup and stuff out of the box. So

28:40

yeah, I do love that

28:42

flexibility and freedom. I did quite a

28:44

bit of Rails and Django myself earlier

28:46

on in my career. And

28:48

it is funny

28:50

how the types of things that you lost

28:53

by moving into this JavaScript framework

28:55

world that we never

28:57

really quite reclaimed in the JavaScript

28:59

world. Like what? Like

29:02

what? Okay, so you

29:04

could think of the closest analogy

29:06

in this, and I'm in JavaScript,

29:08

but in PHP, like a Laravel

29:10

system, right? You get like, so

29:13

you have a templating syntax, you get scaffolding,

29:16

you get migrations,

29:18

you get ORM, you get

29:21

routing. So you

29:23

get more, especially

29:26

more in the data piece that

29:28

you get in JavaScript. That's an

29:31

area that I think the JavaScript frameworks are afraid

29:33

to touch, you know? Oh yeah. You

29:36

know, one thing I really like from

29:38

Rails is the model stuff, like an

29:40

active record, and then all

29:42

the callbacks, like lifecycle callbacks they have.

29:45

Active record, I know ORMs have a little bit of

29:47

a bad name now, but I really like active record.

29:50

It does a pretty good job of giving

29:52

you the 80-20 solution for

29:54

an ORM, and then you can kick out to SQL when you

29:56

want. But it has a lot

29:58

of lifecycle hooks. And it does, I think, a really

30:01

good job of allowing you to

30:03

encapsulate your business logic or

30:06

your domain logic in model

30:08

object and model classes. I

30:11

like object-oriented. I've got a bunch

30:13

of my popular opinions. I like object-oriented programming. But

30:16

you know, one thing you mentioned, Scott, a little earlier was

30:19

people saying, oh, we're coupling,

30:22

with HTMX, you're coupling your front end to your

30:24

back end. And that's true, they're not wrong. There's

30:27

an essay up on the HTMX website called

30:29

Two Approaches to Decoupling. And

30:32

HTMX does couple your front end to

30:35

your back end in

30:37

a pretty dramatic way compared with

30:39

a generic JSON API. We

30:41

actually argue that's a good thing because

30:43

your web app tends to be very

30:45

churny and you need lots of little

30:48

features or whatever. And this can put

30:50

a lot of, if you have a

30:52

generic JSON API that's serving that, then

30:54

that API ends up getting very

30:56

complicated because of that. So you either have

30:58

to introduce something like GraphQL, or you

31:00

have to create a bunch of specialized endpoints for

31:02

your app or whatever. And

31:04

so separating those two actually helps

31:07

your general purpose data

31:09

API be more generic and satisfy

31:12

the people that need that. It

31:14

gets the chaos of your web app sort of

31:16

out of the JSON API. And

31:18

then there's also this idea,

31:21

as we pointed out, HTMX

31:23

does let you pick whatever kind of back

31:26

end you want. And that's because you've decoupled

31:28

sort of at the network architecture layer.

31:30

So you basically decoupled like a level

31:32

down rather than at

31:34

the application level, you've decoupled at the network

31:36

architecture level. And so that is a different

31:39

style of decoupling, but it does give you

31:41

this sort of freedom to pick whatever back

31:43

end you want in a way that you

31:45

can't if you're more committed to

31:47

a tight integration

31:49

between client side JavaScript and

31:51

server side JavaScript. I

31:55

even saw somebody embedding their entire

31:58

API. and

32:00

a worker, it was like a go

32:02

compiled to Wasm through their

32:04

entire API and a worker using

32:07

a SQLite and everything's local on

32:09

device and shipping HTMX. Again,

32:12

this is a wild

32:14

world that we're entering here. Yeah,

32:16

I've seen some pretty crazy mashups

32:19

with HTMX. I can't say I support all of

32:21

them because I don't understand all of them. But

32:26

you know, school, have fun. Yeah.

32:29

How do you feel about web components? I

32:32

haven't used them very much myself. HTMX2

32:35

is going to have better support for them.

32:38

They suffer from what frames, iframes

32:41

suffer from. There's

32:43

almost too much isolation. I don't think they did

32:45

a good job of thinking through

32:47

how to integrate them. There's so much

32:49

focus in the components world on isolation.

32:53

But there are just times when you

32:55

need to integrate. So form participation has

32:57

always been kind of difficult and so

33:00

forth. I

33:04

understand they're improving, but I'm

33:06

not an expert on them. So I don't have

33:08

a super strong opinion. I've never

33:10

needed them. I've always achieved

33:13

similar ideas by using server

33:15

side template reuse, like inclusion and

33:18

so forth on the server side.

33:22

And I think that the firewall that

33:24

kind of got in the heads of

33:26

the designers, we have to have this

33:28

hard firewall between the component and the

33:30

rest of the DOM has hurt

33:32

their adoption. But again,

33:35

I'm not an expert. That's just my

33:37

outsider opinion on it. What

33:41

are the big templating engines that

33:44

people are using? What

33:46

are you reaching for? I like

33:48

any template system that supports something that

33:50

we call template fragments, which

33:52

is the ability to within

33:54

a bigger template to call out a piece

33:57

of that template and say, this

33:59

is a separate piece of content that I want to

34:01

be able to render by itself. That's

34:04

kind of an obscure feature of template languages.

34:08

More templating languages are picking that up as

34:10

they see HTML being

34:12

adopted and hopefully they read that essay

34:15

on it. But what that lets

34:17

you do is it lets you have one page

34:19

with your content in it and then if you

34:21

need to dynamically re-render a part of that page,

34:23

rather than extracting that out and having another file

34:26

to deal with somewhere else and you can't see,

34:28

you can just tag some of the

34:30

content in that template and say, okay,

34:32

I want to be able to render this separately. And

34:35

that works really, really well with HTML because

34:37

then you can say, okay, you know, something

34:39

like if you're doing a poll, for example,

34:41

you can say, okay, just re-render just that

34:43

bit of the screen, but I

34:45

still can see everything in one place. That's

34:48

one thing that I like about

34:50

it is like, you can still sort of see

34:52

everything in one place.

34:54

I'm so glad you said that because like, in

34:57

the past like two months, I've brought up

34:59

the fact that Svelte can't do multiple

35:02

components in a single file and everyone's like, ah,

35:04

just bring them into another file and react server

35:06

components. If it's a client component, you got to

35:08

put in another file. And I've gone so far

35:11

to try to like change my

35:13

editor so that I can edit them

35:15

in line, like put a mini editor

35:17

inside the VS code because like, I just

35:19

want to look at the whole thing

35:21

and type where I want it to go.

35:23

Like it was HTML, but a lot

35:25

of these templating languages make it

35:27

really tricky to do that type of thing. So, and

35:30

I was thinking the other day, I was like, am

35:32

I wrong? Am I wrong?

35:34

And I'm glad finally, somebody else

35:36

is on board with me. Yeah,

35:38

there's another, I'm sorry to keep referring to

35:41

the essays page, but there's another essay up

35:43

there called locality of behavior, where

35:46

we talk about kind of exactly what you're, the

35:49

big idea of what you're talking about, which

35:51

is you don't, and this is a criticism

35:53

I have of the idea that you should

35:56

separate scripting and styling and HTML all apart.

35:58

Is it in order to. understand what's going

36:00

on on a given page and you have

36:02

to look in like 10 different places. It's

36:04

very difficult to you know you have

36:07

to keep a lot in your head to understand everything that's

36:09

going on and so we the terminology

36:11

we try to use around that is locality

36:13

behavior. You want to put the stuff like

36:16

where it's being used as much as

36:18

possible and so I think

36:20

template fragments are sort of an example

36:22

of an implementation where you get more locality

36:24

so you can just see okay on this

36:26

page this is what's going on cool alright

36:28

move on with my life I don't have

36:31

to navigate around so much. I

36:33

saw an interesting thing today object

36:36

oriented HTML is an interesting repo. Have

36:38

you come across that? I just saw

36:40

it for the first time today. Yeah

36:42

I saw it on Hacker News I

36:44

didn't dive into it too deeply but

36:47

I don't know what was your

36:49

take on it did it look interesting? Yeah

36:51

it looked interesting because I think specifically it

36:54

does kind of what we're talking about here

36:56

where in line with your HTML you're able

36:59

to define like okay

37:01

even just in line this is a

37:03

template here this is a fragment here

37:06

in one

37:08

specific file and even

37:10

as far as you could reference another

37:12

HTML file it seems like kind of

37:15

the way that I almost wish

37:17

HTML was invented to be

37:20

where you could have your own fragments and

37:22

templates both in line or imported. Yeah

37:26

it's in a tab somewhere I need to read it. I

37:29

was just reading before this conversation which is why it's not

37:31

on my brain I have no idea if this is a

37:34

valid approach or not or anything but I thought it was

37:36

kind of interesting. Is

37:39

there any kind of local

37:41

first story in HTMLX given that there's

37:44

so much server required this is something

37:46

that I've it's been on my brain

37:48

just a little bit I don't I

37:50

don't know you know I haven't

37:52

been doing too much local first myself.

37:55

Is there anything there that is she

37:58

Mexican can do? Not

38:00

at HTML, it's like, you know, in

38:02

my mind again, if you're talking about

38:04

HTML, you're talking about HTML plus plus.

38:06

And so HTML doesn't have

38:09

much to say about local first. It's

38:11

their hypermedia controls. You can set up

38:13

a web worker, I believe, to

38:16

handle requests. And so you can

38:18

kind of intercept the network abstraction

38:21

a little bit from that sense. And I've seen people

38:23

do that, but it's not

38:25

something I would do unless I was

38:28

really pressed into service to make local

38:30

first happen. So I would say, you know,

38:32

we say this in the essay on when

38:34

to use hypermedia. If local first is a

38:36

must for you, then HTML is probably, or

38:39

hypermedia is probably not a good thing, and

38:41

HTML is probably not a good thing. So...

38:44

Yeah. You've mentioned HTML2 in

38:46

this conversation. I guess that goes

38:48

into one of my questions, which

38:50

was like, what does the future

38:52

of HTML look like, given its

38:56

rise in popularity now? Yeah,

38:58

I think the future of HTML looks very much

39:00

like the present of HTML, which

39:02

I hope looks very much like the past

39:04

of HTML. I think the core, like, you

39:06

know, again, there's seven core attributes.

39:10

The core abstractions of the library

39:12

are correct in as much as they

39:14

implement what my idea of generalizing hypermedia

39:17

controls is. And,

39:19

you know, there's flaws,

39:21

like, if I could take things

39:23

back, I would. But we're not

39:25

planning on breaking backwards compatibility in

39:27

HTML2 very much at

39:29

all. There's going to be a couple of sort of

39:31

corner cases we clean up a little bit, but

39:34

not any major API

39:37

changes. So most HTML1 users

39:39

should be able to upgrade

39:41

HTML2 without any changes

39:43

at all. And my hope

39:45

is that, you know, we're going to

39:47

drop IE support for in two. So

39:49

that'll let us clean up the implementation

39:51

a little bit, use some more modern

39:53

JavaScript features internally. But the

39:56

API should not change. So

39:58

and I don't anticipate it changing. much

40:00

going forward. That's my hope anyways.

40:04

Let's talk about the

40:06

Twitter stuff. We went

40:09

as far as to title our episode,

40:11

is HTMLX a joke? Because

40:14

like props to you, but you're

40:16

like a number one troll online

40:19

and just run the most hilarious Twitter

40:22

account. And so much as to be whenever

40:25

somebody talks about HTMLX, they have to say,

40:28

I know the Twitter account is a little

40:30

bit much, but it's actually a pretty smart

40:32

way to build applications. So how

40:35

did you learn to be such a good troll

40:38

online? Good troll online. How

40:40

did I learn to, I think, shoot

40:42

post is the word you're looking for

40:44

there. Yes. You

40:47

know, I grew up in like on the

40:50

forums in the late 90s and early

40:52

2000s and there was

40:54

a lot of rough and tumble and I

40:56

got pretty good at like, you know, making

40:58

memes and stuff like that. And

41:00

I didn't do it for a while because I was,

41:03

I had a startup going and intercooler

41:05

JS kind of, I don't want to say

41:07

it died, but it definitely flatlined. Like, you

41:09

know, it had nowhere near the attention that

41:11

HTMLX was getting now. And

41:14

when I came back and did HTMLX, I had this

41:16

Twitter account that still had a fair number. You know,

41:18

for me, it was a lot. It was like, you

41:21

know, a couple thousand followers, which seemed like a big

41:23

deal at the time. And I was

41:25

like, you know what? I'm just going to have fun

41:27

again because I kind of, I tried to be professional

41:29

a little bit. I mean, you can go look at

41:31

some of the, you can look at intercooler js.org website.

41:33

There's some silly stuff there. But

41:35

I did try to be a little bit more

41:37

serious about things. And with

41:39

HTMLX, I was kind of like, you know

41:42

what? This is just, this idea is dead

41:44

anyways. I'm just going to have fun and

41:46

run the Twitter account the way I want

41:48

to. And

41:50

I did that and it, you know,

41:52

it just worked out. People, you know,

41:54

there's not a lot of strategy here.

41:57

This is just me being real. fun

42:00

and there have been a couple of times when

42:02

like I you know I got lucky like that

42:04

whole Microsoft you know the whole

42:07

setup for you know me pretending

42:09

like Microsoft was gonna buy out HTMX

42:11

I just got really lucky like

42:14

you know there and so I got

42:16

this stupid email from Bed Bath and

42:18

Beyond from Microsoft so you

42:20

know there have been a couple of times whenever

42:22

you know tried to set things up like that

42:24

but for the most part yeah it's just me

42:27

being me being me so

42:30

can you tell the story about the the

42:32

whole Microsoft thing because I had caught parts

42:34

of it but I I could not piece

42:36

every all the hilarious together and I was

42:38

like I'm not gonna you

42:41

know when you like see like a sub tweet on

42:43

Twitter you're like I don't got time to

42:45

research what this is about what's the background

42:47

here yeah I think if someone had

42:50

someone there was a job posted at

42:52

Microsoft that mentioned HTMX that's what started

42:55

it and so there was a this

42:57

job and it didn't even say like oh you have

42:59

to be an HTMX expert it was like you it

43:01

was you know as a comma like after React

43:03

in view and all this other

43:06

stuff like at the very end they were like

43:08

oh and HTMX and so

43:10

someone retweeted that and I

43:12

was like oh okay let's we'll have some fun

43:14

with this and so I like I

43:16

just I quote tweeted that and said oh this is

43:18

inner I think I just said interesting I tried to

43:21

be like really mysterious about it and

43:23

then I started like finding people

43:25

that were mentioning Microsoft positively and

43:27

like quote tweeting them and saying

43:29

interesting I did that for

43:33

about a week and then literally like this

43:36

is just shows it's way better to be

43:38

lucky then then good I got an email

43:40

from like the Microsoft credit

43:42

card division offering

43:45

me a credit card with 10% off on

43:47

Bed Bath and Beyond and and

43:49

and I wanted to change the license of HTMX from

43:55

BSD 2 to BSD 0

43:57

because BSD 2 requires

43:59

that you distribute the license and copyright

44:01

with it, which doesn't make a lot

44:03

of sense in a JavaScript library. And

44:06

so I started,

44:09

I think on Friday, or I forget

44:11

what it was, but I was like, first

44:14

of all, I made my account super professional looking.

44:16

Like I took out all the junk, I

44:19

made my icon like normal and so

44:21

on and so forth and just started being like

44:23

really serious to freak people out. And then I

44:25

posted, I think on Saturday or Sunday night, I

44:28

was like, okay, I have a very interesting offer

44:33

from Microsoft. And I was just talking about that

44:35

stupid credit card offer that I had gotten. And

44:39

I'm gonna announce a license change

44:41

tomorrow to guarantee the financial stability

44:43

of HTMX going forward. You know,

44:46

just trying to get people freaking

44:48

out, you know, about another

44:50

open source rug pull coming in coming.

44:53

And then the next day I just

44:55

said, okay, it's BSD zero. And the

44:58

offer was this stupid credit card offer. So

45:02

that was probably the biggest setup I've done.

45:05

You know, beyond that, it's all just been

45:07

like chaos. Like, you

45:09

know, at some point I think, I hope

45:11

Elon Musk actually takes the bait on one

45:14

of my tweets and that'll be, I'll probably

45:16

delete my account after that. Oh,

45:18

that's great. That's

45:22

good. Okay, so we're about 50

45:24

minutes here. Is there anything

45:26

that you want to get across about HTMX

45:29

that we haven't covered yet? Yeah,

45:33

I think one thing I've been saying this on most

45:35

of the podcasts I've been going on lately is I

45:37

do like the

45:39

idea of deescalating the language

45:43

around HTMX and just the web development

45:45

in general. HTMX is

45:47

a tool. I think it's a useful tool.

45:49

I think people who are React developers could,

45:52

you know, learn about it and find something

45:54

useful in it and maybe find it useful

45:56

for like an internal project or whatever

45:58

if they've just got to get a button. up that does

46:00

a thing, maybe they don't need all that stuff.

46:03

And the ideas of hypermedia are

46:06

interesting, just from a technical

46:08

standpoint. So I think you

46:10

don't have to, and I can't tell

46:12

people who hate HTMX how to behave,

46:15

but if you like

46:17

HTMX, I would ask that you be reasonably

46:19

balanced in your promises around

46:21

it. It's not a silver

46:23

bullet, and software development is always hard, regardless of

46:26

whatever tool you pick. I've

46:28

been trying to de-escalate stuff because I do

46:31

joke a lot, and so I'm like, God,

46:33

everyone would be nice. And then

46:35

I post some outrageous meme and it's like, wow. But

46:40

in general, I

46:43

think it's an interesting tool which you can learn

46:46

pretty quickly, and there's some interesting ideas behind it.

46:48

Don't feel like it's a

46:50

life or death situation. Yeah,

46:52

well, I've noticed that. I mean, it's

46:54

not anything new, but people

46:56

are so tribal about their

46:59

tools, and their

47:01

tool that they find to be

47:03

effective is apparently the only tool

47:05

that can be effective. And

47:07

I think that's just a wrong place to be in life,

47:10

which is one of the great things about us getting to do

47:12

this show is that, Wes and I,

47:14

we get to try a bunch of different tools. We

47:16

don't have to necessarily deeply

47:19

ingrain them into our personalities. We can talk

47:21

to people like yourself and really get a

47:23

good handle on what these things are good

47:27

for and maybe where they're not appropriate. But

47:29

yeah, I think that's a good advice for anybody

47:32

out there, and not just fans of HTMX is

47:34

that you don't got to be tribal about

47:36

your tools. I think these tools all

47:38

exist for reasons, right? They all exist for

47:40

a reason. And just keep an

47:42

open mind about them. OK. So

47:45

now is the part of the show where we can get

47:47

into sick pics and shameless plugs. Did

47:51

you come to the show with

47:53

a sick pic or a shameless plug? Let's

47:57

see. I'm going to... I'm going to...

48:00

recommend your listeners look at two pieces

48:02

of technology. I will

48:04

recommend they look at Alpine.js. I don't know

48:06

how much you guys have talked about Alpine.js.

48:09

Yeah, we had Caleb on the show. Okay.

48:13

And so that's really interesting. That

48:17

library has a lot of locality

48:21

associated with it. And then another one

48:23

that's a smaller library

48:25

that kind of has a

48:27

mix of HTML and Alpine

48:29

together is something called Datastar.

48:32

And I believe the URL

48:34

is data-star.dev, which is it

48:36

was made by someone who came on the Dave

48:38

Delaney who came on the, he

48:41

came on the HTMLX discord and

48:43

had a bunch of ideas around how

48:45

to improve HTMLX. And we basically were

48:47

like, no, too complicated. So

48:50

he did the right thing. He went off and did it

48:52

himself. And so that's another approach

48:54

to sort of this hypermedia oriented model.

48:57

It's more sophisticated in some

48:59

ways. But I think that's

49:01

like if HTMLX doesn't grab you because it's

49:03

too simple or stupid or whatever, maybe

49:06

go check out datastar.dev. Cool.

49:10

Yeah, I've never heard of this. Yeah, it's

49:12

pretty now. They also

49:14

have an essays page, all these

49:16

academics getting into web development.

49:21

Yep. And shameless plugs. Is there

49:23

anything you would like to plug?

49:27

Um, I guess, you know, my book,

49:29

Hypermedia.systems, you can read it online for

49:31

free. If you want to, you can

49:34

pay for the Kindle version or we

49:36

sell a hardcover version of it as

49:38

well. The hardcover version I've

49:40

heard has had some printing issues. So

49:42

it's unfortunately Amazon on demand is

49:45

apparently somewhat erratic in what it produces.

49:47

But the cover is really cool, which

49:49

is why you're buying the hardcover anyways.

49:53

Yeah, the website is really cool, too. Yeah,

49:56

it is. And then there's also swag.html.org if

49:58

you want to go buy. HTMX stuff

50:00

including there's a poster of the covers if

50:02

you don't want to pay 50

50:04

bucks for the hardcover You just want the

50:07

you just want the cover for a poster

50:09

We saw the the smallest version of that

50:11

poster at cost because Berkeley graphics,

50:13

which is now I think us

50:15

graphics He changed his name They

50:18

did an unbelievable job on the cover and

50:20

so we wanted to make that available to

50:22

people they want There's a lot of

50:24

stuff on swag, too It's

50:27

really cool. What's up with the pickle? What's

50:31

up with the pickle? Someone generates that pickle and

50:33

I took it and hacked it up a little

50:35

bit and Photoshop or I use Photo

50:38

P. I think is what it's called the online Version

50:43

and and then I use

50:45

it. I don't know man. I don't know what's

50:47

going on It's just

50:49

really really funny and like awkward

50:52

and sort of very HTMX Yeah

50:56

It just it just works. I don't know what to say about it

50:59

There's a lot of good vibe stuff on here like the

51:01

the platter pickle It's like a platypus and

51:03

the buffalo with that laser eyes and you

51:06

know the influence. Yeah, we try to have fun You know

51:11

You might as well have fun. You might not be successful. You

51:13

might as well have fun Yeah,

51:15

well, it's been awesome Carson

51:18

It's been really great getting to talk to you and

51:21

I hope the audience has a little bit better of

51:23

an idea about the intricacies of HTMX

51:25

after this one and Yeah,

51:28

it's something that I personally been really

51:30

enjoying working with so I'm gonna continue

51:32

to dive deep into this world and

51:35

I'll see ya online great. Well, hey, thank you

51:37

very much for having me on I appreciate it

51:40

Thank you, you know just

51:42

just again. Thanks for taking a look

51:44

at it and admittedly idiosyncratic

51:47

Way to do web development Love

51:49

it. Cool. Well, thanks so much Carson you