[Kde-games-devel] KSokoban: Who will revive it?

Ian Wadham ianw2 at optusnet.com.au
Fri Dec 21 20:49:05 CET 2007


On Fri, 21 Dec 2007 07:29 pm, Dmitry Suzdalev wrote:
> On Friday 21 December 2007 00:53:24 Ian Wadham wrote:
> > So, are you still interested in KSokoban? ... and is Dmitry?
> > Let's sort this out.
> Well, currently I have no plans and motivation on porting&fixing KSokoban.
> Or it will be more correct to say "If I will do it, then this will happen
> in future - i.e. not *now*"
>
I am thinking someone would need to start soon if it is to be
revived for KDE 4.1.  Maybe next month?

> > So, if anyone is going to do a re-write, please at least scrutinise the
> > old code thoroughly and try to understand it fully, however difficult
> > that may be ... and it is &@*!^% difficult in the case of KSokoban.
>
> Unfortunately, I would say that in KSokoban's current state rewrite is the
> only good option. Have you looked at the code, Ian?
>
Yes, I have.  I was looking at reviving KSokoban for KDE 4.0, but decided
KJumpingCube was more achievable in the time frame available last June.

> It's terrible IMO. non-const references everywhere, C++ classes written in
> a C-mode (you know, that happens when one writes in C++, but isn't
> using it's features or using them to a very small degree), no good class
> hierarchy, mess in level saving/loading etc.
>
> These memories come from the time I last looked at KSokoban, so I might
> have something said not correct, but I surely recall that general picture
> was *uuuuh, what an ugly code!*
>
<smiles> Style, aesthetics and purity of language do not bother me so
much as general unreadability, lack of comments and obscure choices
of data names.  I daresay I am guilty of writing a fair bit of C++ in C-mode,
being an old C programmer from way back ... :-)

However, if it is hard to understand the code, it is extremely difficult to
maintain it, enhance it and upgrade it to new versions of KDE.

> What I am saying is that if  I was to improve KSokoban, I wouldn't go for
> any other option than rewrite :)
>
I tend to agree.  But if you do not know what is in the old code, how do you
know what to write?  And how do you avoid leaving features out accidentally?
I do not see a detailed specification or design document for KSokoban
anywhere ... do you? ... :-)

All the best, Ian W.


More information about the kde-games-devel mailing list