The future of the DDTSS
We're severals years on and it looks like the "DDTP/DDTSS will
eventually be replaced by Pootle" is as far away as it ever was. In
the meantime the current system, while it works, is not ideal. It runs
on a Berkeley DB version 1 DB which has suffered some damage in the
past and now contains rubbish here and there. Also there have been
very few contributions. So I have some ideas about the future:
1. Switch the backend from BDB1 to the postgresql instance on the
machine. I've actually prototyped this and it appears to work. I just
put a shim BDB interface on a table in the postgres database. It's
ugly and evil, but it would solve the data corruption/list issues. It
would probably then get appropriately backed up. Performance won't be
great but I doubt it's any worse than what is there now.
Unless there are objections I want to try doing this in the weekend.
2. Long term it would be good if there were more people who could work
on the system. I have the feeling that the current codebase, being
essentially Perl with not very good abstractions reduces the number of
potential contributors. If I were to do it again today I'd build it in
Python using Django. My feeling is that this would make it more
accessible for new people. Does anybody have any ideas about that?
The nicest thing is that rebuilding it would allow us to remove
several layers of abstraction, the current system being built to send
and receive emails from the DDTP. With direct access to the DB several
new features become much easier. Perhaps the web interface for the
DDTP could be included making improvements there as well (it could use
a facelift).
The problem with rebuilding is that you run the risk that it never
gets done. On the other hand I think the current interface is a
dead-end development-wise. Note I'm not talking about the backend
scripts, just the frontend and the DDTSS.
Any comments?
--
Martijn van Oosterhout <kleptog@gmail.com> http://svana.org/kleptog/
Reply to: