3 ms·
Same thing happens in software development at the library level: "Why not use [Example OSS Library]?" "Nah, I can do it in 15 minutes." "You know, they have
by ErikAugust 10y ago
Same thing happens in software development at the library level:
"Why not use [Example OSS Library]?"
"Nah, I can do it in 15 minutes."
"You know, they have [X] contributors, lots of great documentation, and 6 months of work into it."
"Yeah, I know [but whatever.]"
- jkchu 10y agoI get the point that you are making, but if something really only takes 15 minutes to do, I wouldn't necessarily use an open source solution just because one exists. Depending on the situation, the library may have a ton of extra bloat that your application will never use. Also, recklessly deferring to open source libraries can potentially snowball into something like the left-pad fiasco[1]. [1]: http://blog.npmjs.org/post/141577284765/kik-left-pad-and-npm http://blog.npmjs.org/post/141577284765/kik-left-pad-and-npm
- ErikAugust 10y agoYeah, and I get your point too. Eight lines of code? Yeah, you don't need a library. But usually when someone says "15 minutes" they mean 1 - 2 days. So things of that complexity is when it's obviously you should choose a good OSS lib.
- hex13 10y agoIt depends. In many cases you would be right, in many other cases you wouldn't. It depends on many factors: - what is the problem? (some problems are really hard, some problems are trivial and just seem hard) - is developer experienced enough to build own solution? - are existing solutions good or badly written?(I often refrained from using existing solution when I looked on bloated code in it). - is solution suitable to project? ("too much bundle size", "doesn't support X", "not too configurable", "too slow", "too hard to use"). etc.