Re: Scripting

[Chris <chris@starforge.co.uk>]

Jan Ekholm wrote:


>>interpreted solution may well be "simpler" - easier to understand, or less 
>>syntactically complex, but it is often easier for me to write the more hairy
>>version because it lets me do things in a few lines that would otherwise 
>>need many.

> Ok, you may be right there. Could you give some examples of things that
> are much simpler in a compiled language?


To start with, anything involving remotely "unsafe"* memory tricks or type 
fiddling - complicated pointer work, interleaving types in memory blocks, 
custom memory handling, that sort of thing: situations where the type 
checking and controls an interpreted language imposes prevent the "simpler" 
or more powerful solution from being usable and require additional code to
solve the problem in a different way (which gets you both ways - you end up
doing more work in a language that is slower, so your program is slower and
bigger than it need be..)


>>You've never really gone near the innards of a fast 3D game have you? They 
>>tends to be very far from static inside - there are all manner of routines 
>>applied to the underlying map before you ever get near creating the data to 
>>send to the card. Grab a copy of the Quake2 engine source and have a look
>>at it, some of the stuff going on in there will surprise you. Or have a look
>>at some of the stuff on www.vterrain.org, especially the LOD stuff...

> No, I haven't. How many of the games out there actually are "fast 3D"  
> games?


Quite a lot - if not with Linux versions. But does that matter? Read many of
the aritcles on gamasutra, check out the more advanced stuff on gamedev.net
or opengl.org and you'll see that things like scene generation, especially 
if you're doing fancy effects like particle systems or doing physics 
modelling, complex collision detection and so on, are not trivial problems.



> So they are on this list? So what? So they are Tetris clones? So what? The
> issue is IMHO not about that Linux game developers do simple games, but
> what language they should be given for their Tetris clone. Recommending
> the new developers to automatically go with C++ is just a little bit
> silly.


Why? In work I'm teaching bioinformatics students C and Java so they can 
write sequencing programs, database frontends and the other weird and
fiddly tools they need. They start with simple problems and work through 
progressively harder ones - why is it so bad to start making small games
in c++ and then go on to bigger ones?

 

Chris
-- 
  .------{ http://www.starforge.co.uk }-----. .--------------------------.
=[     Explorer2260, Designer and Coder     \=\ P: TexMaker, ROACH, site \
=[___You_will_obey_your_corporate_masters___]==[ Stack: EETmTmTRRSS------ ]