[Evolution] Review of GMail.

[Not Zed <notzed ximian com>]

http://www.extremetech.com/article2/0,1558,1586090,00.asp

This was off slash dot.

Anyway, 2 interesting things:
 - threaded flattened conversations with hidden quotations (dunno how badly that works with inline quoting), which was mentioned somewhere just last week.  Also dunno how it copes with multiple branches of conversation; which would be tricky to make sense of if it isn't done right.
 - the way their 'vfolders' work, which are based on keyword tags.

The last is kind of how I was thinking evolution vFolders could have done things like moving individual messages to vFolders, rather than having to add it to the search criteria.  And if you do this for all vFolder operations it makes your vFolders a hell of a lot faster since the query becomes the much simpler 'is this tag set'.  Much much simpler.  However, the problem is things like renaming folders becomes a lot more expensive since you have to go and rename all the tags too (or just use id's, but then you're lost if something goes wrong and you lose the id->name mapping).  And the other problem with evo is it has multiple folders to store the data in in the first place, which makes things like moving vs copying 'tricky'.  e.g. if you move, it only sets a tag, but then it sets it deleted.  So it appears deleted, and expunge removes it, which is definitely not what you wanted (obviously you then disable move).  They also have the advantage of controlling their store, rather than going through say IMAP, which may or may not have the relevent features required on the backend (e.g. storing keywords on messages).  But the real problem is trying to fully support the folder metaphor, I guess, which they don't do on purpose.

Anyway, I guess there's a couple of possible future ideas there.

Michael Zucchi <notzed ximian com> Ximian Evolution and Free Software Developer

Novell, Inc.