[Pan-users] Re: PAN project's status

[Duncan]

Grzegorz Wiktorowski posted <address@hidden>, excerpted below,
 on Wed, 18 Aug 2004 08:45:01 +0200:

> Is PAN project dead or can we expect any new release ?

Well, they say open source projects are never dead as long as there are
users and a copy of the source still around.  However, it does sort of
look like PAN's at least sleeping rather soundly, and perhaps in a coma,
doesn't it?

However..  Even before pan went virtually dormant, there had been
something holding up major progress for some time, tho there were always
other things to do.  Many of the long planned features such as automated
multi-server handling have been being put off until PAN's data handling
was rewritten to take advantage of a data-base backend library such as
sqlite, much like the networking rewrite to take advantage of gnet.  One
of the problems was that Charles wasn't really comfortable with database
work and had no experience in the area.

The good thing about this mainline dormancy is that a number of folks have
taken the opportunity to start implementing various database stuff. 
Currently, they are doing much of this on their own, tho it does happen to
be the big item of discussion on the devel list.  However, mainline
dormancy helps by making it less of a moving target.  By the time Charles
/does/ get time to work on PAN again, or decides that isn't going to work
and hands off his maintainership to someone willing to pick up the reins,
it's possible one or more of these database experiments will be mature
enough to move with quite fast, and PAN development will take a quick jump
ahead, gaining the database and possibly additional features quite
rapidly, where if it had continued as it was, they may not have been added.

Of course, the biggest problem with PAN at this point is its lack of
scaling, the speed and memory problems that occur when one tries to use it
in groups with a million overviews (aka headers), or with a couple gigs of
cache.  Even if the database wasn't going to be used for additional
features, PAN's growth is somewhat stunted until that can be taken care
of, and a database is generally considered a well proven approach to this
sort of problem.  Thus, here again, this "resting" may end up being a good
thing, if it means PAN gets the database backend when things start moving
again.

All my opinion, of course..

-- 
Duncan - List replies preferred.   No HTML msgs.
"They that can give up essential liberty to obtain a little
temporary safety, deserve neither liberty nor safety." --
Benjamin Franklin