Many libraries provide features that were unavailable in packages that are small enough for comfort. Those features and sizes vary per project, and, as admitted, sometimes there's great use for htmx. It's just not going to be for front end, local-first, SPAs. As the author, himself, said very clearly.
Invokers don't give you what htmx gives you out of the box. If you want to, for example, have an invoker command to fetch a resource and swap it in to the page, you have to write custom JS for that. That's the thing that htmx gives you out of the box, along with error handling, progress indicator support, history support, animated transitions, and a host of other features.
If you don't need or want all these and are happy with the worse UX, then of course you don't need htmx.
Re your suggestion: datalist makes you select an option, then fills in an input with the value of the option. Then you have to submit the form to actually get the resource you just searched for. Active search gives you the links to the resources directly in the search results, so you can load them with one click. There's a big UX difference.
As you should be able to tell from the other comments, this is not a library for me. If I had ever looked into it past trying to actually implement it, I would have realized that it's actually specifically for server-rendered content and using it for front-end development doesn't actually add much. The author made that very clear in the supporting documentation, I just never bothered to check.
So I relayed my personal experience with it and then, because of this great comment section, I became aware that I was trying to use it for things that it isn't helpful for. Because, yes, I would have to write custom JS for the functions I would try to invoke with invokers, but I would also have to write custom JS for the functions I would try to invoke with htmx. There's no change, for me, other than including 14kb that I do not need to run a PWA.
So you can take whatever issue you want with my impatience and my myopia in disregarding it as not useful, but my complaints are not immaterial. It doesn't do what I said it doesn't do, and you seem pretty happy to confirm that. I shouldn't have expected a cat to bark, and I regret it. Hell, I'd apologize if someone felt mislead, or if someone feels I'm misleading people in my aspersions. But - like any app - my apps swap out html all of the time and there's nothing that htmx provides that helps me with that in a way that's worth the weight.
> it's actually specifically for server-rendered content and using it for front-end development doesn't actually add much
It's for server-rendered content that can be used to build a frontend app. I know because I've done it several times.
> I would also have to write custom JS for the functions I would try to invoke with htmx
The difference is with invokers you would have to re-implement everything from scratch and with htmx you typically only need to implement some parts that it doesn't handle, like eg listening to a DOM event and doing some action in response.
iOS Safara user here. The viewport jumped for me only on the active search demo and only the first time when the result content was initially loaded. This can be easily fixed by having a fixed-size results box loaded from the beginning.
reply