this doesn’t worry me, though. those in the react crowd that insist on this arbitrary and newfangled “frontend/backend” stratification and are dogmatic about it are by definition going to stick with what they know and won’t come and bother us who choose tools based on real experience and their practical merits. better off they make themselves easy to spot from a distance.
took until maybe 2015 before i remember seeing any job ads of “frontend” positions, it takes a while before a job market develops around a technology.
i didn’t say it was pointless. i said the stratification of “dev” into “frontend dev” and “backend dev” is newfangled and arbitrary. you could also split devs into other classes (DB only, CSS only, etc.).
it is funny when people who don’t know you accuse you of incompetence because you don’t like their tools or methods. dogmatic. i prefer some other tools and methods and i am delivering value to customers. pragmatic.
our policy is that for widgets that are like browser components e.g. search as you type with keyboard shortcuts, we just use the off the shelf react component for that purpose and use it from htmx like it’a browser input element. for all other business logic (almost all of which has no low latency requirements and almost always involves a server requets), we use htmx in our server side language of choice.
our designer who knows a bit of react is not happy, but the 12 engineers on our team who are experts in $backend_lang and who are tired of debugging react race conditions, cache errors, TypeScript front end exceptions, js library churn, serialisation bugs, etc. are very happy indeed.
it doesn’t fit every app, but it fits our app like a glove and many others that I’ve considered writing that I didn’t feel like bothering to do so before discovering htmx.