Basically, shared objects and dynamic libraries are just the same. You might refer to the difference between a .so that is loaded when the program is loaded (e.g. like libc, and like virtually all .DLLs in Windows) and one that is loaded at runtime using dlopen() (just check man dlopen), which is the equivalent of a Windows LoadLibrary().

If you can''t decide between Windows and Linux, just do both. It''s _really_ easy with cross-platform libraries like SDL (www.libsdl.org), and there''s also a Windows port of Qt available for GUI applications (Qt is just the best C++ toolkit out there).

It''s a sad truth that Linux is too "chaotic" to succeed as a desktop OS. Which isn''t too surprising - after all, the actual Linux is just the kernel + some closely related tools (module tools come to mind). Everything else is GNU software and then the bloated X when it comes to desktops.
OSs like AtheOS (www.atheos.cx) have a much higher chance of becoming relevant desktop OSs.

BTW, Linux _originally_ wasn''t written as a server OS. It was written as a free Unix clone for home computers, fulfilling an actually more DOS-like duty
Of course this changed, and nowadays its really strong sides show in the server area.

So Linux isn''t perfect for desktops, and for normal users Windows is better. I still use Linux instead of Windows for the most part, but that''s a personal decision: I just believe operating systems should be free, and I vote with my money and support.

cu,
Prefect

One line of sourcecode says more than a thousand words.

I''ll be bold saying that linux and opengl will not be major market players in the near future. Linux is for people who like the hobby like it was during altair days. Opengl is in trouble and 3DLabs(I think) is trying to head v2 api specs but if that fails goodbye opengl and hello dx9. Seems winxp/dx is gaining lead. Remember the last thing you want to do is dink around with the os/drivers when you''re trying to concentrate on coding your app and deadlines are looming ahead. The money is in windows apps like Carmack and others have said over and over again.

Unfortunately, I think JKD is right

I took a look at 3dlabs proposed specs for OpenGL 2, and there seems to be a lot of problems with OpenGL1.3. It seems as if Microsft is heading in the right direction with DirectX in terms of offering features, and making it more user friendly.

I think we are seeing the disdvantages of having a democratic board making decisions. They are being choked by IP decisions and the reviewal process of the ARB. I think someone needs to step up to the plate to effectively be a "leader among equals" and take OpenGL by the throat and guide it along. Whether it be 3dlabs, SGI, Nvidia or ATI, I don''t really care as long as someone can make OpenGL evolve faster than it has.

This one aspect alone is making me consider programming for Windows more than anything else, as if DirectX9 gets better than I don''t see much of a gaming future for Linux. It can and will be a hobby, but nothing serious.

OpenGL is cross-platform, and that becomes extremely important once you consider applications beyond gaming: CAD, CAM, CAE, etc. For these types of applications, desktop computers just don''t (currently) provide the kind of power and speed necessary, so they resort to high-end workstations. And as long as there are high-end workstations, there''ll be OpenGL.

Now as to its developing too slowly, well - sorry. I never went with OpenGL. I saw quite early the value of having one entity make design decisions on a general-purpose API in terms of rapid repsonse to user requests/demands and emerging technologies. So what if they break the API every time? Just define D3D_SDK_VERSION and so forth and your old apps compile with the new libraries. If you look closely at Microsoft you''ll see that they throw designs out the window (no pun intended) pretty frequently: Win3.x gave way to Win9x and a whole new interface paradigm for the PC. Win9x has given way to NT kernel-based systems, of which XP is the latest incarnation (once again featuring a radical interface change). That way MS has been able to respond to consumers and give them the latest features.

DirectX has hit version 8.0, and 9.0 is in the works. OpenGL, which predates it, is struggling to hit version 2.0. And don''t say that there''ve been minor version numbers; 1.x only has 4 (.0, .1, .2, .3).

In the larger context we see the same problem in free software: GCC 3.0 only just starts to attempt to comply with the C++ standard defined in November of 97! How do your users begin to enjoy the benefits of advancement if you can''t bring it to them quickly enough? That''s why there''s still a viable business selling compilers for UNIX: GCC doesn''t cut it in the cut-throat commercial sector. (Disclaimer: Do not misconstrue this as a lack of respect for GCC; the compiler suite is one of the best out there - reasonably fast, a good optimizer and generating solid code. I merely speak of the rate of implementation of features considered crucial to modern development.)

Hi,

Regarding gcc - Thats one of the disadvantages of open source and free software, you have to wait or modify it to suit your standards, but also it was a boon to those who couldn''t afford a regular good compiler and hence they could go ahead and give us some decent application.

Regarding Opengl :-
I don''t think Opengl will go anywhere - maybe just out of the gaming world.

From what I have seen Microsoft seems to hate Linux - I don''t really know if they do - so Direct3d port etc probably won''t happen in the future.

This are just my opinions.

quote: Original post by flame_warrior
From what I have seen Microsoft seems to hate Linux - I don''t really know if they do - so Direct3d port etc probably won''t happen in the future.

From what I''ve seen, heard and experienced first hand, it''s largely justified. I''ve seen the most rabid anti-Microsoftism amongst Linux supporters, often without basis or reason. The problem with this is that it distances the general public from Linux, because the general public is accepting of (or resigned to) Windows and these "hackers" are telling them they''re morons for using it. If someone consistently called you the "Evil Empire", you might hate them too.

This is not to abdicate MS of all wrong; they''ve done a lot of underhanded bullying to ensure their market dominance, but the vast majority of Linux supporters can''t clearly articulate what exactly they hold against Microsoft.

History: Microsoft once tried to bring UNIX to IBM PC compatibles; their system was called XENIX. Microsoft still develops BSD software.

Dauntless, you''ve said it much better

It might be the case that in the future we will have windows as the only modern os to do our work on and that includes games as well as CAD and other specialized apps. Remember that video card is extremely important so if no one is making one to work with alternative oses then no apps are written for those oses. Today, gaming cards have workstation gfx card features and are becoming faster and faster. I wouldn''t be surprised if businesses are replacing old computers with specialized gfx cards with gaming cards and windows2000 or winxp. What''s the need for platform independence if 99.9% of customers are using windows os?

Few years ago there were many os players and through time only the strongest prevail i.e. microsoft. This always happens, first a flood of companies fight for market share then only handful of them survive and go on. There is a chance to undercut the price of a product of large corporation and allow newcomer to enter but how often that happens when large corporation has established contacts (has lots of customers) and buys in larger quantities where price is discounted is not known to me. I think this is true for video card makers as well.

Businesses need to stay on top of their competitors and need fast support and updates to their apps they bought from companies. I wonder how many commercial products are created with linux gcc, probably not many compared to msvc++. Software companies that use 3rd party libs/dev tools need reliable supporter who doesn''t go bankrupt or doesn''t disband for lack of user community members. These dev companies do not need source code they need the code to work and that''s it. It cost them to hire someone to look over the 3rd party source code and improve it. Then when someone else makes updates to 3rd party open source code the code becomes incompatible with the other company making improvements to it unless both users of source code work together but then that would mean giving out secrets and position of upper hand. Not a good idea for a business that wants to lead the market don''t you think?

JD, I think you''ve touched on several important issues for the Linux community. Commercial entities don''t need source code integration; for them binary compatibility (.lib, .dll, .exe) is sufficient to build their specialized solutions. This is a major reason why Linux'' fragmentation is derisive: too much software must be compiled from scratch on every single machine ! (RPM and apt notwithstanding).

Furthermore, platform independence is a moot point when there is only one major or profitable platform. And the software industry is really just mirroring the PC hardware industry. Initially there were huge profits to be made because there were few companies, but over time the competition killed off lots of PC makers (Zeos, Olivetti, Tandem...) and now the few that remain seem to be gradually consolidating. The pickings are slim. Microsoft has wedged itself into the number one place, making support of other platforms an inefficient proposition. And as you pointed out, that''s just business.

Whew, what a voluble way to say "I agree."

Hmmm, JD and Oluseyi

These are all good points, and as much as I wish it weren''t so, I think I need to focus my time and concentration with Windows instead of Linux as a programming platform. Maybe in time once I get familiar with Windows I''ll tackle Linux, but with the way things look right now, I don''t see much opportunity (or future) for Linux gaming.

I don''t like M$''s business strategies, but I have to admit, they''ve made a more compelling platform, and they have the resources in place which allows for easier implementation. As JD said, there''s not much use for platform independence if everyone uses the same platform (although it WOULD be neat to have direct comparisons of people playing the same game on PPC and x86 systems).

In a lot of ways, I think the Mac fans have more to look forward to than Linux fans in terms of gaming. At least they have a standardized setup that programmers would have to rely on (and hey, OS X is sort of a cousin to Linux anyways). In my tiny scratching of the surface of Linux programming, it really confused me how the setup process was done (i.e use make to maek a makefile that you run with ./configure to check for system requirements, then autoconf...blah blah blah).

That being said, if Linux could standardize itself better, and had a more centralized set of tools and infrastructure to build on, then I think Linux would be a very viable gaming platform. But I''m not sure if Linux (and OpenGL/SDL) are heading in this direction. Also, despite having said that I think Windows is the better GAMING platform, Linux does seem very interesting for beginning programmers (despite its diffuculty). I think Linux forces a beginning programmer like myself into the more nitty gritty details of coding, and things aren''t as "black boxish" as it is in Windows. And thanks to it being open source, beginners like me have a better chance of figuring out how and why certain things work.

So I really would like to get back to Linux sometime, but for now, I think I''ll stick with Windows (I''ve always been a generalist that knows a little bit about everything, but is good at nothing...I need to stick to windows until I get good at it, then I''ll branch out).

I look at SDL and at the games loki''s made. I cannot imagine how people can sit there and say "Linux isn''t good for games." Linux is an OS. Windows is an OS. I can take Loki''s code and compile it against SDL on windows. (Almost from scratch.. I would imagine SOME things would need to be done.) Likewise, I can take a game I am developing in SDL and compile it on windows.

Cross platform compatibility will dictate whether games will be or won''t be made for linux. I plan on coding with SDL. My games will work on Linux. They will also work on windows.

To me, that means there is no such thing as an OS gap for my games. Discussing whether the OS will live or die on the desktop is pointless to me. It is a good OS, and it can be made to do whatever developers wish. No, it''s not the best in the world. No, I don''t think it will ever beat Windows. But it is a viable option, and coding for it CAN be done.


Magnwa