| I still don't consider myself a veteran but I saw the web evolves from 2010 up til now. And imho we took a wrong direction. Web apps and pages have become so bloated and complex that's crazy. I'm just starting to recover from my previous work. I had to maintain migrate and add features to a legacy system (built in 2017) which had initially a GraphQL api and a SPA, but that was later split in 12 microservices (with cycling dependencies because hey why not) and 3 Big react applications, all of that in Javascript + Typescript (added later with all the gradual typing stuff that makes you think that your code is correct). All of that to serve 300 monthly users between 35 and 50 years old, who just cared about filling forms here and then and have some charts over data. 0 rocket science. So far I remember history going like this:
- 2010 Clean SSR with rock solid boring technologies and some vanilla JS where it made sense, I remember page load where crazy fast and you could not feel the need for an SPA.
- 2012 NodeJS arrives. It came to save us from slow blocking IO apparently (??)
- 2016 React arrives. It came to save us from boring old SSR apps
- 2017 All React + SPA frameworks (in house often)
- 2017+ Someone has the brillant idea to do SSR with React. "It's a revolution"
- 2017 GraphQL arrives. You are saved.
- 2020 Nextjs arrives. You can now do static generation for SEO and SSR and apis ! That's great. Thank you Next !
- 2022 Someone wants now to do Server Component that will make your page load going insanely fast (which we already did in 2010 when we carefully loaded vanilla JS at the time) I learned a bit of ASP.net core recently, and I think that boring razor pages are still relevant. I remember my first lead in 2010 telling me that ORMs and SPA where completely overrated. I didn't agree with him at that time. Now 12 years later I think he might have been correct. I'm not saying SPA have not their place. Some apps can't do without it. I'm saying that we are victims of "hype driven development" which involve having some trendy hashtags in your resume and convince your manager that you have seen an incredible new tech that will make the business incredibly wealthy.
Except it does not and you end up with a legacy system that nobody wants to maintain because too complex. Maybe like Jonathan Blow stated in one of his videos, we make a confusion between going forward and progress.
Now all the new kids learn React in some boot camp or online but have no idea how to implement a map function. And I feel old and grumpy. And tired by all this. |
I never understood why the answer to the "frontend problem" was to do more things on the frontend.
Writing yourself and API to talk to your own backend is silly nonsense, yet somehow it has become the norm. I loathe it.
Thank science for Phoenix LiveView and the various frameworks that followed in its footsteps. Also htmx and the like.