Re: libmikmod woes..

[Steve Baker <sjbaker1@airmail.net>]

Christian Reiniger wrote:
>
> >The shame of all this is that there was/is an effort to build one
> >STANDARD game library (The Linux GSDK project) - I joined that at
> >the outset - but they wasted so much time discussing and discussing
> >what should be done that I gave up working with them and went my
> >own way.

> ...now the PenguinPlay as you know it is officially dead.

> But as giving up is too easy - and plain wrong - there's now a new
> PenguinPlay. Same basic goal (creating/defining a standard game SDK), but
> completely different means of archieving it. I hope I can have the proper
> new homepage up until this evening.
> 
> Steve (I knew I knew you...)

...damn - is there no escape?!   :-)

> can you help with it? I mean, not the "Uh, we have a plan,
> can you do the work please" thing, but can you have a look at
> PenguinPlay from time to time and give me a slap when you think
> it's progressing too slowly, when there's too much talking again,
> and too little action? And if I then come up with some stupid
> excuse, slap again twice as hard?

Well, I think that what went wrong with Linux-GSDK (Long before you
were involved IIRC - and certainly before it had the PenguinPlay
name) was that it needed to get something basic WRITTEN more
quickly. Something concrete for people to criticize and build
upon. When Stephane started GSDK, he wanted it all finished
in six months - a very smart move!  However, there was no
interest in building on existing technologies - which is particularly
silly when you have such an ambitious goal and such a short
timescale.

You can't run a huge cooperative freeware project like you
would a commercial project.  Writing specs up-front, planning
everything out and then moving on to the coding just isn't
the way to go.  Talk is cheap and when 95% of the subscibers
to your list will never contribute a line of code, you can't
afford to bog down the 5% who will contribute with the endless
talk that goes on when 100 people all try to influence the
work. Dive in, write code - be prepared for it to be rewritten.
Establish who those 5% of productive individuals actually are.
Weigh your decisions accordingly.  When someone who is not
an active contributor says you are doing it all wrong, just
say "OK - contribute something better and we'll use it" and
99% of the time you'll never hear from them again.

When GSDK started (many years ago of course), it was a great idea
- there was then a real need for some kind of games SDK for Linux...
but times have changed. There are now a bazillion games SDK's
kicking around on the web.  Starting up GSDK again at this late
stage seems rather pointless.  You'll never again have the
chance to make a single GSDK that everyone could sign up to,
and any new initiative would just add yet one more library
so we'll have a bazillion-and-one games SDK's and even more
fragmentation of effort.

If PenguinPlay does restart, I'd be happy to listen in on
proceedings - but you have to realise that from my perspective,
the problem is largely solved.  I have PLIB - which is exactly
what I wanted from GSDK.  I'm now into the maintenance
phase and looking to start serious games writing (and I'm
already 50% of the way into my first game - which has already
been downloaded about 15,000 times - by any real measure, that's
a success).

PLIB's isn't "finished" - it's scope could certainly be extended
- (eg 2D graphics, etc) - but I don't ever see the need to start
all over again with a completely new SDK. I certainly don't want
to fight those same battles all over again.

I'm sure you'll hear the same story from the authors of
ClanLib, CrystalSpace, etc.

I'd advise anyone interested in restarting the GSDK/PenguinPlay
project to pick the existing SDK that comes closest to their
personal goals (eg 3D, 2D, high-level-genre-specific,
low-level-general-purpose) - and join an existing team. They
all need help.

Better still, don't write a new SDK at all. Write a game
instead. It's much more fun - and Linux desperately needs
*new* games. The existing SDK's are easily good enough to
let you do that.

It would be nice to think that all the developers of other
libraries (myself included) would say "Oh Wow!  At last a
single package to unite all these little SDK's together -
I'll drop everything and work for PenguinPlay." ... but that's
*NOT* going to happen.

GSDK is a lost opportunity - let's give it up and move on.

-- 
Steve Baker                (817)619-2657 (Vox/Vox-Mail)
Raytheon Systems Inc.      (817)619-2466 (Fax)
Work: sjbaker@hti.com      http://www.hti.com
Home: sjbaker1@airmail.net http://web2.airmail.net/sjbaker1