lgonze at panix.com
Fri Jan 9 11:37:15 PST 2004
I'm in the process of writing a module to guess the mime type of a URL.
To do this I need a list mapping extensions to mime types. Writing
that myself that would be dumb, so I surfed around, but I didn't find
anything really good. The IANA registry is terrible, completely out of
date, and even if it weren't it is not formatted in a way amenable to
generating a map.
I found a pretty good mime.types file at my ISP, though it lacks a
number of mime types that matter. The best code I found was:
I also found a good algorithm at:
However the mime.types from my ISP is still very incomplete; it doesn't
have .m3u or .ram, for example, and it also lacks the metadata needed
So, well, this is a stupid situation. Obviously what I have to do is
manually add missing extensions as I find them. But this should really
be a standardized list maintained by a group. Does that exist already?
An alternative to a manually generated map is to use crawler results to
auto-generate one, perhaps with human editing of the results.
More information about the FoRK