[rycfan]

If I engage in as much hyperbole as the author, where does "write it yourself" stop? If I'm working on a team of two, should we each write our own left-pad? How about a team of three? Four? Five? Fifty? At a certain point, it makes sense for that to be written once for the project. We spent 30 years in software engineering trying to figure out how to get code re-use, and now that's it common and widespread, we want to go back to NIH?

[marssaxman]

I think about this every time I write a UTF-8 decoder. It's a task which is simple enough that I can bang it out from scratch about as fast as I can find a library that does it, so I can't really stomach adding a full-blown dependency to deal with it - but it's also just complex enough that I have to think about it, write some tests to make sure I didn't forget something, and wonder whether this implementation is any different from the last one I wrote.

[kbart]

I have 'tools' repository on my local gitolite server just for that. Saved a lot of time already.

[sealjam]

I pretty much came to the comments to post something similar. I'm sure there was a long comment thread here on HN discussing the perils of NIH this week. This seems like the end point of the alternative.