https://publicobject.com/2025/01/05/try-something-simpler/
=> More informations about this toot | More toots from jessewilson@cosocial.ca
@jessewilson Good quote pull!
I was amused that you used the word "hubris" in this post, since your "this should never happen, so crash so that we can find the file/line later" is the approach we tend to use in Hubris. 🙂 Armstrong's philosophy has had a lot of influence on my software.
(I assume that ! generates a file/line so you can find it easily.)
=> More informations about this toot | More toots from cliffle@hachyderm.io
@cliffle yes!
My first draft of the post directly claimed that fail-fast is correct and that the never-crash idiom is wrong.
But I decided to post with a bit more humility.
=> More informations about this toot | More toots from jessewilson@cosocial.ca
@jessewilson I feel like there's a place for both in an application. Like, an application container / webserver should not crash, but I want its individual request serving tasks to fail fast. You can compose the two arbitrarily.
=> More informations about this toot | More toots from cliffle@hachyderm.io This content has been proxied by September (3851b).Proxy Information
text/gemini