Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

This person would disagree about point 4: https://www.kalzumeus.com/2010/06/17/falsehoods-programmers-...


Author here! Ha, I'm a big follower of Patrick, specially his microconfs! I didn't know he had an article an article on the subject, thanks for linking it!

However I agree with him!

What happens is that in my experience I've yet to encouter a case where what I suggest in the blog doesn't work. No doubt that there will be cases where it will not be enough. But by reading Patrick's article, clearly the complexity of the system needed to handle all the cases would be huge! So I guess one has to compromise :)


Yes, compromise in that typically US-centric way, and to hell with any culture that doesn't fit the mold, they can adapt!

It's true, they can adapt! They proved that when they didn't have much choice, back in the ASCII days when you had to conform or stay in the pre-computer age. But maybe we should try to better now that we can...


So, my identity records got mixed up due in part to the impedance mismatch between (first name, last name) and (first name, middle name, last name), resulting in my legal name suddenly changing until I could get the issue sorted out.

Two fields works in the sense that most people are going to be have something to put in them, but what that is and what they mean isn't always going to match what you intend. Either the system is going to be tolerant of such things (in which case you can probably get away with some variant of (display_name[, external_name]), or you're using it to interface with some other identity provider/consumer that requires a specific format, and that's what should be asked for.

Names are personal and hard to get right.


Very, very few cases aren't handled by an arbitrary utf8 string of not more than say, 1KB.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: