Thanks everyone!

Summarizing available "thesisable" todo projects you wrote:

1. Multiple server/database connections from within a single sql
statement. (Dalton Calford)

-For me to understand this problem it would need some more specification
and I would need a bigger understanding of how Firebird and it's
transactional mechanism works.
-If not implemented/designed yet it really IS Sexy. I'd guess this would
take quite a bit of research and not less work to achieve something.
-Big enterprise deployments of firebird would benefit from this one



2. Design a database connection driver for mobile devices. (Ruben
Carballo Macklis)

-For example something like a Firebird JDBC driver for mobile phones?
Sounds interesting and probably would have it's uses.
-Seeing that everything is converging to mobility this is a step into
the future :)



3. Open-source free and production ready replication solution. (Roman
Rokytskyy)

-Now this is something I understand, and without further research know
it's uses :). I guess a lot of people (myself included) would appreciate
a good cross-platform solution. Haven't taken a look at Alan's
FBReplicator yet, nor any other existing solutions for Firebird though.
I like the idea of "physical" replication. Nickolay has already
implemented such a solution, but what do we know about it?
-Everyone could make use of this one



4. Securing the database and the client-server communications. (Doug
Chamberlin)

-Very important and nice feature. After securing it, I'd add optimizing
client-server communications for the Internet too if that's possible.
This is a feature that every user would benefit from either small or big
enterprise. Something to add to Firebird's already superb palette.




A bit of a fable below, don't read if you haven't got time.

I'm currently in a dilemma. This being my last year means I am somewhat
bound to my consultants and to Ericsson. Seeing my hesitation they have
come up with something vaguely interesting: tcp on high bandwidth and/or
high delayed (rtt) networks. The current TCP implementation (~TCP Reno)
doesn't work well so I've got to focus on and evaluate related work
(HighSpeedTCP FastTCP etc) in wireless environments.

Doing a thesis on Firebird has loads of very appealing positives, one of
the biggest being doing something that actually helps progress and gets
used (the problem with the whole so-called scientific and research area
is that it's really 99% bogus, nothing to do with real practice). On the
other hand it involves digging deep into firebird, of which I hardly
have any knowledge. Having lots of experienced people and equipments
directly at hand makes staying at Ericsson's idea much more comfortable.

But I know, that If I'd have come across helping Firebird a year or half
a year ago I wouldn't be thinking about a comfortable thesis, of what I
can be rid of as fast as I can.

At the Budapest Technical University we have about 2 years beforehand to
do research on something that will probably be our Thesis. I myself got
into an unfortunate situation where 3 of us did the research and only
one got to do the Thesis. I'd guess other universities do it the same
way. So if you're in a relationship with a Uni, you can get quite a lot
of help on the project as long as you give active supervision and help
to the students. Were there any similar cases before?

I also like the idea of Ubuntu's bounties:
https://wiki.ubuntu.com/BreezyBounties

Cheers,
Sean
--
"It is often easier to not do something dumb than it is to do something
smart."