Hacker News new | ask | show | jobs
by marginalia_nu 607 days ago
To be fair, the Last-Modified header is very sketchy. I use it as one of the heuristics for determining the age of a website in my search engine. It's not great.

It's frequently found incorrect, both older and younger than the actual age of the document. It's a bit of a relic from back in ye olden days when websites were static .htm files in a folder, which is so rarely the case today.

It doesn't help it's also got overloaded uses via If-Modified-Since -style conditional requests.

1 comments

Plenty of websites also play with the user-visible dates on websites to game search engines - most dates shown in Google results seem to be complete garbage. I don't think Modified-Since is really worse, and it at least gives you a chance to maybe get a date for static pages.

But you are right that If-Modified-Since forces it to be a date for the complete document rather than the content, which might not be as useful to normal users for dynamic pages.

Yeah, my takeaway after having attempted to do so is that properly dating websites is a very hard problem. You can get Google-level-accuracy decent guesstimates relatively easily, but going beyond that is hard.