💾 Archived View for tilde.pink › ~kaction › log › 2023-11-02.1.gmi captured on 2024-07-09 at 00:24:34. Gemini links have been rewritten to link to archived content
⬅️ Previous capture (2023-11-14)
-=-=-=-=-=-=-
Today I stumbled on the blog post titled "Software disenchantement", which talks about topics close to anybody who would read by gemlog -- software bload and software sustainability. I can't agree more, I try to do it right for my personal projects that I build in the copious free time, but let me tell you story of my $dayjob.
https://tonsky.me/blog/disenchantment/
I am personally responsible for gradually building a monster that takes five (!) minutes to shove 50Mb into several Postgres tables. I know how to make it faster, but it was never a priority. What is a priority is to pile more and more hacks that would handle more and more malformed data. JSON schema and HTTP 400? Referencial integrity? You wish! xkcd#1179 is a child play. Why so? Humans.
Software project that do not try to make money can tell humans to RTFM or get lost. Business can't do this. Either /you/ cater to humans, or somebody else does, and /you/ are now out of business. And you'd better adapt to whims of humans fast, or again, you are out of business. Nobody cares about performance or memory usage or other stuff that you can buy from AWS for $50/month, but whims must be granted ASAP.
The "Software disenchantement" post ends on positive note:
As engineers, we can, and should, and will do better.
I would love to believe, but no. Whenever money are at stake, you can sacrifice neither scope or nor time for the any quality above "acceptable".
If it is possible to make money out of X, somebody will do X. This is why we can't have world where software development is done exclusively by unpaid volunteers: it is possible to make money out of software development.
And you know, 40 hours a week is a lot of resources, so majority of software /will/ be produced by businesses and, ergo, suck.