Re: [TuxKart-devel] Slowness - really related with Karts?

[James G. <j....@vi...>]

On Mon, 2004-09-06 at 02:02, Steve Baker wrote:
> James Gregory wrote:
> 
> > Because you've brought it up again, I'll put the case forward again. No
> > dependencies are better than dependencies, sure. But given the choice
> > between a crap GUI and no fullscreen mode, or a nice GUI and fullscreen
> > mode but with dependencies, or a nice GUI and fullscreen mode and no
> > dependencies but a couple of months of extra coding, which is best?
> 
> There is NO indication that you can't have a decent GUI with the PUI toolkit
> that TuxKart was originally using. 

It depends if by "decent" you mean "functional" or "looks like a
computer game rather than a multi-coloured office application".

>  I explained that fullscreen mode should
> be obtainable with PW - but if it couldn't we could modify it to do so (since
> I also run the PLIB project - this should be easy).

You never seemed like you were going to do anything about it. When SDL
was originally put in it was guarded by #ifdefs, giving you a chance to
say "OK, I'll sort out fullscreen with pw if you remove SDL". Instead
you just responded with a flame. Our response was to say "Well, if
you're going to be like that" and give up on trying to find a middle
ground.

> 
> I wasn't enthusiastic about expanding the scope of the GOTM effort to encompass
> a drastically re-engineered GUI - and I think (given the extension of the project
> beyond the originally estimated 2 months) that I was correct.  

Most of the basic GUI code was converted from Neverball to TuxKart in
one evening, with maybe another evening's worth of my time spent
improving it further. Where myself or other people have used the gui to
write e.g. character select screens, it has been no more effort than if
these screens were built using plib. So the only person who has put any
sizeable amount of time into rewriting the GUI of tuxkart when using
plib could have freed their time for other things is me.

However, as I have freely admitted on multiple occasions, I don't know
very much about OpenGL, and so it's not like I would have fixed e.g. the
falling-through-the-track bug instead. The main obstacle to a release in
the immediate future remains the 3D engine, yet this isn't something I
could have done anything about. You could have, but you decided you'd
rather see GOTM fail (which isn't going to happen, but it is delaying it
hugely and it is going to mean it won't be as good as it could have
been).

You are being totally irrational in arguing that the reason for this
GOTM going on so long is because I spent a couple of evenings switching
over from plib for the gui. The real thing behind all of this is that
you wrote PLIB and you'd like to see as many games as possible based on
it. You are outraged that a version of a game that you originally did
most of the work for now uses rival libraries instead. I guess in a
world where people are motivated by the credit they gain (i.e. OSS),
this sort of thing isn't surprising (RMS does it plenty too), but I
still think it's stupid.

> However, *had* I
> been interested in doing that, PUI could have done the job perfectly well without
> introducing more dependancies.

Again, your definition of "perfectly well" is not shared by everyone.

> 
> > As I said already, the use of ttf for in game text was a matter of what
> > was easiest, not what is best.
> 
> What would have been easiest would have been to use PLIB's font library and not
> getting into having to cache fonts and all that nonsense.

	James