[Kde-games-devel] Possible GSoC project

[Guy Van den Broeck]

2008/3/24, Ian Wadham <ianw2 at optusnet.com.au>:
>
>
> Looking at the CSPoker website, it appears that CSPoker is still at
> a fairly early stage in its life-cycle (first Beta) and has already had
> quite
> a long development period.  There is nothing wrong with that.  It is
> quite normal and to be expected.  But what I ask is whether it is
> realistic
> for Guy to expect to complete very much within the time-frame of a
> GSoC project?  Might you be biting off more than you can chew, Guy?


That depends. If I have to rewrite all the server logic in C++ then yes,
that would be too much.
If I can write a thin client for the Java server (possibly hosted on GGZ
Gaming Zone?) then there's no problem and I can even throw in some basic AI
for a single player mode. My Master's thesis is about poker AI.
I understand your hesitation but I just finished a JavaFX thin client in a
few weeks while studying.  I'm sure I can do a lot more work this summer. I
don't want to rewrite things from scratch.
I'm not sure what a long development period is but we've only started
CSPoker this summer and I'm pretty happy with the stage we're in already. In
a few months we should have a very stable fully functional game.

I think KDE 4 Games definitely needs a Poker game, but to produce a
> really good one, engineered to KDE standards, could take quite a lot
> of effort and involve several people.  If we were to start such a project,
> what would be the best starting point?  KPoker, CSPoker or something
> else?  How do we decide?


I think we should compare features, what is under the hood and how
> easy/hard it would be to build upon either KPoker or CSPoker, as
> opposed to starting from scratch.
>
> For example, KPoker only plays 1 or 2 players, is not networked, plays
> only standard 5-card draw-poker, has an AI but perhaps not a good
> one, would like to be able to play Texas Hold 'em, has nice KDE Games
> card decks which have been taken over into KDE 4 and (KPoker) is
> already a KDE application, written in C++.

CSPoker plays any number of players, is networked (but perhaps not yet
> reliably), plays only Texas Hold 'em, has no AI (yet) but might have robot
> players, would like to play other kinds of poker, has only one card
> deck(?),
> but nicer graphics otherwise, and (CSPoker) is written in Java, is in a
> Beta
> state and is not a KDE application (yet).


With all due respect to KPoker but other than a hand evaluator, there's very
little to it. The work we did on CSPoker to build a scalable, multithreaded,
multiplayer server is of high quality and well engineered. I also think that
our card deck looks very good compared to KPoker's.

> BTW do KDE applications really *have* to be written in C++?  Whether
> they have to or not, I know from personal experience that it is quite a
> lot
> of work to KDE-ize an application.


This indeed is a key issue. I can write a thin client in C++ but if we need
to bundle the server then Java needs to be supported.

Please correct me if I am wrong about any of the above and I am sorry
> if it appears negative (I would say it is realistic ;-)).  I would really
> like
> to see a revival of poker in KDE 4, but I think it is more than a short-
> term project.


I appreciate your honest evaluation. I only want to do this if the community
supports it, and if they do, I want to work very hard to achieve the goals.
CSPoker is a long-term project for me and that will hopefully include a KDE
port.

Cheers,

Guy
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.kde.org/pipermail/kde-games-devel/attachments/20080324/4be750e1/attachment-0001.html