Mime header parsing...

[William Shipley]

We do in fact understand the MIME header just fine.  However, right now we don't do a heck of a lot with  
it; we let the underlying library handle it.  In fact, with the library version we are using, we can't seem to  
get at the header directly; we just have to trick the library into doing our will with the file types we want.

That's why we haven't done anything with file sizes; it's in the header, but the version of the library we  
have actually doesn't understand it and treats it as an error.

We will probably extend our handling of file types to deal with more of them inside OmniWeb (.Z, .gz,  
and .rtf ones are high on the list), instead of letting the Workspace do as it wishes with them.  Right now,  
we only attempt to handle images and sounds inside OmniWeb; everything else is just passed to the  
Workspace.

We're trying to coordinate with CERN to update the WWW library and get our fixes rolled in.  Up until  
now we've basically been trying to avoid modifying the library at all, so that we could benifit from future  
releases.  (Unlike Mosaic, which has totally modified the WWW library.)

-W