Re: Choosing a Linux development platform

[Christian Reiniger <warewolf@mayn.de>]

Rick Genter wrote:

>1) How similar are the various Linux's (Linuxii?) right now? I plan to
>support Intel Linux under Red Hat as well as LinuxPPC. If I do development
>under LinuxPPC, how portable will my efforts be to Red Hat Intel (machine
>architecture differences aside)?

The main differences are in the location of config files, startup scripts
and (to a lesser extend) libraries. But that should be pretty irrelevat to
you. Another problem might be different library versions (mainly
libc5/libc6), but that's not too difficult to overcome either - a
relatively simple installation routine can handle that nicely.

>2) The last time I looked at Linux (about a year ago), there was a
>Win95-like window manager that worked reasonably well. Is there a
>Finder-like window manager available for Linux?

IIRC it's named mlwm (Mac-Like WindowManager). But you're most likely
better off with KDE or GNOME. Both have about the same functionality as the
Mac desktop and can be configured to look and behave as one.

>3) Are there GUI IDEs available for Linux (other than Codewarrior, which I
>am aware of)? Commercial, shareware and freeware leads are all welcome.

Code Crusader is a free one inspired very much by codewarrior.
gIde is in a rather early stage.
You could check www.linuxprogramming.com for some more.

But - you should seriously consider using a good text
editor+make+autoconf+gdb/ddd
No IDE can be as versatile as that combination. And with a larger project
both options (IDE / many tools) cost about the same time (The multiple
tools require some more learning while the IDE will waste some of your time
because it can't handle some stituations as you'd like it).

>4) What is the current state of gcc with respect to: a) C++ language
>support, b) optimization proficiency, and c) stability? Note that fast
>compiles are NOT of interest to me as much as correct compiles and compiled
>binaries that run fast.

Well, gcc pretty much means egcs currently. And that compiler is *really*
good. C++ language support is almost complete (the standard library still
isn't 100% complete, but besides that there's no serious flaws).
Optimization - well, I don't know. But I guess it's pretty good (The
maximum optimization level for egcs is 6 (compared to Gcc2.7's 2) and
optimizations from pgss (specially Pentium-optimizing) are integrated
constantly).
Stability - well, all Linux programs are compiled with gcc, and they are
regarded as more stable than their commercial counterparts.

>5) Though as a developer who takes pride in his work I find a certain
>attractiveness to producing software under GPL/LGPL, I don't see how I can
>make any money at it, and money is *a* motivation for me (obviously not the

As someone else already mentioned, Selling the *data* should be the way to
go. That has some benefits:

* The data is taking up all that space on the game CDs, that means copying
it is about as hard as copying an entire game (you can't have copy
protection mechanisms in the code however if you release the source ;)

* Releasing the game source is very gamer-friendly. No more waiting for
bugfix patches. The game engine is constantly being improved, ported to
other platforms and maintained. Just imagine you buy a game for your PII
system and five years later you still can play it - on your new SGI
VirtualStation!
It will live longer and through that constant improvements it can become
more feature-righ, more capable, more versatile than any closed source one.

But you should prepare yourself well for the open source development (I
added some useful links to the Project Management doc (is it updated,
Paul?). Have a look at them).


	Christian
--

Drive A: not responding...Formatting C: instead