Currently there are some ideas emerging for a rewrite that I'll
probably do one
day. Especially it aims towards an improvement in GUI structure. As
you know,
the old Gorm wasn't very well-structured, and articles like [1]
inspired me to
rethink my program architecture, which is currently not very modular.
Another thing that still gets me concerned is that Grr reads the whole
article
and feed database file on startup. Basically, when it closes, it
serializes everything
it got into one large database file, which is then read again on
startup. I'm afraid
it might slow down startup when more and more articles accumulate.
(After all,
noboby wants to throw away articles.)
I wonder if it's possible to store the articles on the disk and load
them on-demand
while keeping the ability of searching them with Lucene.
One way I came up with that makes the transition easy is to keep the
current concept
of writing everything to that database file *but* leaving out the
actual article contents,
which should be the biggest part of that database file. So only
meta-data is stored in
the database file.
Those article contents could then be stored elsewhere (in separate
files) and could
be loaded on-demand when the getter for the article content is called.
Have you taken a look at Vienna?
http://www.opencommunity.co.uk/vienna2.html
It's a really nice feedreader -- there might be some code you can pull
from it.
J.
_______________________________________________
Etoile-dev mailing list
[email protected]
https://mail.gna.org/listinfo/etoile-dev
