gumroad essay

[slavingia]

No description provided.

[wedow]

This sounds like a very interesting development experience. Would you be open to elaborating on your points with any extra details or maybe specific examples?

1. What made the htmx experience feel "unnatural and forced"?
2. What sorts of user experiences were impractical with htmx?
3. What led you to attempt to "implement more sophisticated interactions and state management" with htmx?

There are many use cases where htmx is not a good fit but also many cases where teams have made the opposite choice of moving from React/Next.js to htmx and could copy/paste your list of reasons (aside from AI use) to describe their own experiences. In my own case, htmx enabled significant UX improvements and a much more rapid rate of development.

It would be great to have further insight into how htmx didn't fit your use case so that others with similar needs can relate and make more informed choices for their own projects.

[slavingia]

Could you show your UX improvements in screenshots?

[wedow]

Could you show yours? That's kind of the point. It's also odd to respond with a "no u" when asked for details.

Our use case involved rendering large data sets and we had to contort React to handle it with decent performance but browsers can easily render thousands of HTML elements without issue. Speed and reliability are important UX concerns for us, they may not be for you. The interface otherwise maintained 1:1 parity with the React version so there was no degradation of functionality.

Worth noting that more complex interactions still require some JavaScript with htmx — but it's radically simpler to augment HTML with some dozens of lines of JS as needed than to build and ship an entire JS app.

[slavingia]

I did? it's in the commits, or did I mess something up?

[slavingia]

i added one more example and tried to elaborate to address your comments/wants! lmk what you think :D

[slavingia]

improved it quite a bit imo. hope it's ready!