Many of the ideas and concepts in htmx are things we worked on starting around 2012 at a top tier investment bank. Implementation details differ quite significantly, but the idea of hypermedia driven applications was core to everything we did.

We unfortunately weren't able to win hearts and minds around these concepts in the long run, and blog driven development (a.k.a. cargo culting) replaced our efforts. I feel somewhat redeemed now that HTMX seems to increase in popularity, it's nice to see that we weren't alone in thinking along these concepts.

Of course, if the concepts were strong that implies my execution wasn't, so maybe I shouldn't feel so good about it after all... :o)

Or your marketing might not have been the best possibly, or people didn't spend enough time with alternative approaches to appreciate their downsides and why hypermedia based solutions could be better. (Just pointing out other possibilities because I don't know if it's a good idea to be hard on oneself when one isn't sure about what went wrong.)

These were internally developed tools, so nothing we really marketed externally but yes we definitely could have done a better job at our marketing with the organization. A lot of the criticism to our approach was typically not so much that the ideas were bad, but that they were "different from how everyone else does it" and by everyone else they typically meant Facebook and Google.

For example, global state management approaches like Redux were more or less incompatible with how we approached state management. We thought it was insane to build apps around the idea of global state stores (and I still do) but Redux became all the rage and eventually won the day. There was a lot of collateral damage as people then tried to shoehorn Redux everywhere. I left around that time, but hear that the cycle have repeated a few times since. Redux (and other tooling) didn't work out as expected so wheels have been invented again over and over. It's the way things go I guess, I don't dwell too much on it.

We probably could've done a better job at communicating why our approach was better, but ultimately we had limited ability to force anyone to follow our approach and use our tools. There's also a hiring aspect in there. As Redux and other libraries that were "the thing" started to make their way into job descriptions it became harder and harder to convince top talent that they should join us, because we didn't have that brand recognition and used in-house tooling. At some point you have to make trade offs and sadly top tier financial institutions aren't necessarily known for leading the pack in terms of innovation. It's kind of sad really.

Anyway I'm not particularly hard on myself or bitter about it at all actually. The self deprecating jab about poor execution should be read as tounge-in-cheek really, I'm very proud of the work we did back then.

This was a time when people were shouting "separation of concerns" without really understanding what that meant. Mixing the html/css/js required to make up one discreet UI concern? Heresy! Requiring at least three files to describe a discreet UI concern? TheOneTrueWay!

I remember the issue to be also that managing UI state on the server side was a load issue that many places didn’t want to deal with (e.g. Wicket, etc). The vibe at the time JS client apps came out in my circles was "thank God we can get rid of this UI stuff."

That’s still an issue I have with HTMX, but I understand for simple use cases it’s fine. I’d rather use HTMX than JQuery that’s for sure.

Or it's because you weren't Facebook.

Like angular, react got the huge initial boost in popularity because of the brand behind it.

Absolutely, brand recognition is hugely important in whether something's a success. It can also go the other way, great ideas and tooling simply die because the brand is tainted.

Not everything Microsoft has produced is terrible, but for a while there it seemed all developers I knew couldn't see past the brand even if the tooling/products/ideas were solid. I feel it's gotten better since Nadella took the reins, but the smell still seem to linger sometimes.

I agree - I'm a greybeard, so it's a bit of a bizarre world to me where Microsoft is actually a lovely, valued open source brand and probably the most ethical (depending on your definition) of all the big tech companies these days.

Time to go snow skiing in hell, I guess. It's deeply frozen over.

I'd give Microsoft more of a chance, but the way updates and the UI are handled in Windows really puts a sour taste in my mouth. I've heard many good things about stuff like WSL2, PowerShell, and other things that can make Windows feel more accessible for power users, but I'm still looking at a Windows 7 machine without a TPM. I'll be able to update to W10, but after that I'm hosed and have no choice but to 'upgrade' to a chip that has a nanny CPU in it.

If Microsoft provided a Developer's Edition of Windows -- without telemetry and ads -- that also came with access to a knowledge-base of best practices for Windows-based software, and why you want to use each major Windows component, for a flat, one-time fee, I'd strongly consider building things for that environment.

I feel like Microsoft's current approach of buying their way under the free software world via GitHub is kinda scummy, and all that code's now under Copilot and other LLMs. If they were a bit more upstanding and had a clean, mutually respectful business model, I might consider buying their products.

In the mean time, I'd rather use something written by a random hacker, because I can at least read their code, or even hit them up and have a conversation about their project.

That's not true, see for example initial announcement on HN: https://news.ycombinator.com/item?id=5789055

React didn't get much adoption initially and FB didn't spend money on marketing their framework, like Google did with Angular. Developers started using React to speed-up table rendering in Angular application, and only after some performance-related blogposts adoption started to grow.

> blog driven development (a.k.a. cargo culting) replaced our efforts

Ironic considering what the OP attributes to HTMX's current popularity.

Fair enough. :o)

Nah, look at it this way:

A good idea or a good product are never enough. You have to sell the product in order to have it be successful.

In your situation, you were at a bank, so your TAM was roughly 1 enterprise (or maybe a dozen, tops, if divisions had federated or siloed tech). And your 1-to-a-handful of potential customers were probably of the large and difficult to change variety. That's a huge constraint, and not landing it there doesn't mean your idea was bad, it just means you weren't successful in selling it and I'm going to tell you that you probably _couldn't_ have been successful in selling it where you were, when you were there. I worked for your lazy cousins at the same time - big insurance - and had the same issues selling my own ideas internally, if that makes you feel any better.

HTMX has no such constraints. It's a decade later and the TAM is more or less everyone doing or preferring to do SSR. The same/similar good idea, but different place and time.

I hear you're buying a synthesizer and an arpeggiator and are throwing your computer out the window because you want to make something real.

I hear that you and your band have sold your guitars and bought turntables.

I hear you rewrote your http endpoints to return JSON because that was REST.

I hear that you and your band have sold your turntables and bought guitars.

I hear you rewrote your http endpoints to return templated html fragments because that was HATEOAS.

I'm losing (regaining/losing/regaining) my edge

Is this some kind of metaphor?

An appropriation of James Murphy to make a common potshot about webtech trends being as circular as fashion / music ones

Haha, software has more seasons than the fashion industry. If you observe it for long enough everything seems somewhat derivative. But I wouldnt want it any other way. All diversions are welcomed within the choices of those working within the choices of those before us.

BUT HAVE YOU SEEN MY TECH STACK?

I suspect VC distorts value in tech choices.

I'm pitching a mobile app built in Unity and a custom ruby backend on digital ocean.

They questioned my tech choices on a working app! Apparently if you want VC money you must use their approved tech stacks.

> I suspect VC distorts value in tech choices.

I agree. I think what we're seeing is investment risk assessment as a proxy for a deep understanding in technology. After all, if you don't know how anything works, all you can do is look to what _has_ worked for you and other investors in the past.

I'm guessing here, but it's plausible that once a given stack has proven to make money, the fix is in. It wouldn't matter if it's the best choice out there or not. The track record itself makes it more attractive. Example: Ruby on Rails.

This is a fun semantic game. You can also tell them it uses a Von Neumann architecture and also 8 bit bytes. Or how about it uses Linux, or the latest Linux 6.5-rc6?

Is this supposed to be Rockstar? The reference Rockstar Satriani interpreter barfs on this input.

No, honestly I've never heard of satriani.

But I do appreciate that you tried to parse it.

It's some lyrics from an LCD Soundsystem track with some topical snark mixed in

I heard that you have a backup Data8 tape of every Unix release...

Every seminal bsd

I enjoyed building webapps with Backbone, it was rich web applications that felt like true apps instead of trying to make parts of a site more dynamic using jquery. It was a clean separation between front- and back-end.

But Backbone was a bit sloppy, not very enterprisey, until Angular came around.

Anyway, this string of web applications became popular (in my world) because it meant that we could separate back-end from front-end, and have one back-end (usually REST/JSON) fueling mobile and web clients separately. That was it, that was why we built SPA's, but by now that has been forgotten again.

And in my world, it made sense because the apps were behind a login. But then They wanted to make SPA's facing the internet, for things like webshops. And I get that too; I've made a few websites using Gatsby, they're crazy fast for navigating AND still indexed by search engines.

Anyway, in some cases it did get more and more complicated, then with server-side React and the like. I never had to touch that, thankfully.

React is very diverse community, who are these react bros?

I went to a series of JavaScript conferences and React is extremely diverse.

I was recounting the past. I'm sure it is "diverse" ( whatever that means ) now because it became kind of a industry standard for front-ends.

Good to see diversity all around, I just wish there also would be diversity of thought and it comes to building web pages. But I guess that would lead to people admitting ~80% of all the stuff they build does not require React or equivalent js frameworks and that would be both an ego and economic/résumé problem..

It's an industry standard for people who like React. It's really that simple. This idea that they are so dominant and everyone is forced to use it is complete hogwash. JQuery and Wordpress still dominate the web completely.

Maybe you can explain this to my otherwise reasonable employer.

> ( whatever that means )

I inferred that the intent was to say that not all people using React fit into one category.

> ~80% of all the stuff they build does not require React or equivalent js frameworks

I'm well aware of this for some portion of the things I build, but I don't agree that these frameworks are generally unnecessary or not preferable. It's also hard to say when something "requires" these tools or not. Arguably you could say nothing requires them these days.

I like to build with all kinds of tools though, and React and Vue are for example are often more helpful than harmful to my development experience. I'm speaking as someone who has been building things for the internet for around 20 years.

> that would be both an ego and economic/résumé problem

Maybe. I think this might be more true in roles that are closer to entry level, though.

Do you think that the frontend crowd similarly thinks that you build dead simple 5 page websites with a Rube Goldberg setup of back-end frameworks?

I certainly do.

:) I appreciate you being a long time user and I certainly do a lot of s-word posting on the ol' twitters

on the other hand, my hope is that eventually htmx (and, more generally, hypermedia) is accepted as a tool, a useful too, but just a tool, even by the front-end folks who haven't thought very much about hypermedia in a while

i don't see the two approaches as mutually exclusive and agree w/ Rich Harris' concept of 'Transitional' web applications that mix the two approaches. I would just draw the line where you abandon hypermedia for a more elaborate client side approach in a different place than him.