Sure; but npm packages are much more likely to still be around in a few years than some random startup’s web app. Even the original developers can’t take their own code off npm in most cases any more. And if you really needed to replace a package you depend on, most useful packages on npm have a bunch of good-enough alternatives you can switch to with a little bit of effort.
Re:The web’s most important decision
by shanebellone on Monday May 01, 2023 @12:27PM
(#35775140)
You're absolutely right. Something suitable will probably be available at any future point in time.
Come to think of it, I have been bitten by something like this before. I used the (mostly) single-developer project clojure-android to build an app and had no good options when bit-rot set in.
Sure but even then, forking & fixing that one library will be a walk in the park compared to remaking Figma or Google Docs if those sites ever go down. They don't seem like remotely comparable risks.
Re:The web’s most important decision
by shanebellone on Tuesday May 02, 2023 @01:24PM
(#35791077)
"fixing that one library will be a walk in the park compared to remaking Figma or Google Docs if those sites ever go down"
Most core tech is trivial to rebuild for a single user. It becomes complex over time due to scale and with the addition of user-guardrails.
I've been running a homegrown Object-Oriented Database for months without issue. It's less than 100 lines of Python including the interface.
All code presents risk. With modern dependencies chains, those risks compound through time too.