Re: [glob2-devel] We must do a game, not an engine

[donkyhotay]

I have to agree with bradley on this one. The glob2 code is very
confusing and messy. I'll admit to being one of those that is learning C
++ by reading/tweaking the code. I know much of what I implemented while
working is probably confusing to others while there have been times I
have been unable to complete what I'm attempting to do as well. Anything
that makes the code easier to go through I'm in favor of. 


On Wed, 2007-04-25 at 23:04 -0400, Bradley Arsenault wrote:
> On 4/25/07, Cyrille Dunant <address@hidden> wrote:
> > Well, yes, nice people stepped up after I complained that we would still 
> > need
> > the maps.
> >
> > > Well exept you.
> >
> > This is incorrect. I complain that people keep subtly insulting nct and
> > nuage's work, calling their code a mess and saying essentially that it need
> > to be redone from scratch.
> 
> Well first of all, from reading the code, I've noticed that there once
> was a simple, and usable design. But things have changed. People have
> hacked in new features. There are new requirements on code that wasn't
> designed to meet these requirements.
> 
> And as such, the new features that have been put there have
> increasingly complexified the code. Coupled with the lack of good
> documentation, I have every right to say the code is a mess. I am a
> very critical person, but I offer reason why its a mess, and a more
> maintainable alternative.
> 
> nct and nuage where learning C++ while doing glob2. I don't blame
> them, and I don't name them when I say the code has become bad. And
> there is allot more agreeable design in newer code they've wrote.
> Infact there is a clear difference in style the code that was written
> a long time ago and hasn't changed since, and the code that is brand
> new.
> 
> The problem is that code that was written a long time ago with
> primitive C++ design is well primitive C++ design, and it hasn't
> changed. Instead new code has been written around it, and the
> constantly changing requirements on old code that hasn't changed means
> Glob2's code has become quite messy.
> 
> > I am trying to attract your attention to the fact that if you are here, 
> > today,
> > contributing to glob, it is because we have a working codebase, and not a
> > worthless pile of junk.
> 
> You have a furnace. It works. But its really old. No person in town
> knows how to fix it. It works for now. If it breaks though, you have
> no furnace, and you can't repair it.
> 
> 
> > More than that, we have a game that is fun to play.
> >
> > > I don't reply of the rest of the mail.
> > > There was no information in the rest of your mail anyway - only a lot
> > > of aggression.
> >
> > I ressented a lot the remark that if the maps had been done once, they could
> > be redone.
> >
> > The understatement was that obviously this is basically not technical stuff,
> > and therefore not really important.
> >
> > A good map is good after a few hours of drawing, testing, redrawing,
> > retesting.
> >
> > A good strategy game is good because it has many _perfect_ maps. Think
> > Starcraft.
> >
> > So yes, I am angry, because I have the feeling that non-coding contributions
> > are considered trivial.
> >
> > > And I don't understand how anyone here could be the cause of it.
> >
> > Because you are in essence saying that all that was before needs to be 
> > redone.
> > Ergo, it is crap.
> 
> It has become crap. It was once good. But it has been slaughtered in
> time. It needs to be redone. Requirements change. But the old code
> hasn't.
> 
> > Note, that you are not saying that you would like this or that new feature,
> > no, you say that the code is a mess. You have just described it as "refusing
> > to clean up your room, only much worse."
> >
> > So yes, I am angry, because of the understatement that all previous (old) 
> > code
> > is crap.
> >
> > > I was about to congratulate us of the fine and hard work this month.
> > > Doing the releases, tests, bugfixing, wiki-rewriting, animations and
> >
> > The releases highlighted massive problems in the process (well, better 
> > during
> > the alphas than later on). So yes, I suppose we could congratulate ourselves
> > for having had a couple releases :)
> >
> > And the animations have usability problems (well, 50% alpha could be a fix).
> >
> > > repositorial change while some core rewriting takes place is incredible.
> >
> > You should thank nct for doing that on your behalf. I am not certain that it
> > will improve much the development process, but if it creates goodwill, it
> > probably was worth it.
> >
> > > And don't forget the two servers we changed.
> > > Hopefully this will not again turn out to become a flame thread.
> >
> > It _is_ a flame thread. And would you please stop trying to extinguish
> > perfectly good flames.
> >
> > --
> > -- Cyrille Dunant
> > -- EPFL-IMX-LMC
> 
>