Linux, Opensource, Games
Ok, this may be totally inaccurate (I am still fairly new to linux), but it''s just my opinions and thoughts.
When I first started using linux, I discovered the nightmare hidden in little .RPM files. Rpm dependancy errors left me sifting the web for the (literraly) 100s of broken dependancies layered one on top of another of little libs and tools to install the main program. I then realized that instead of installing RPMs, I could download the source, and do the ./Config, Make, Maike -install dance, and everything worked great (for the most part).
The obvious is that everything in linux is released as opensource, free for modifications and changes.
Now, if you are a game developer writing a client for an online game, do you ~really~ want to release the client sourcecode? I think not. Even for non-online games, I highly doubt developers would want to spend millions on a game, then release all the source graphics, music, sound, and code they worked hard for to eager hands that will most likely steal them and use them as their own.
Correct me if I''m wrong, but I think you''d see more willingness to develope commercially for linux if companies could write programs that were closed source, and could be easily installed on the system..
Feel free to flame me
"Mommy, where do microprocessors come from?"
I will put what he CmndrM said in another set of words.
The idea of linux is to provide a very good open source OS to anyone that wants to use it and provide that source for free. That cought on with the programmers so they will provide the program and source to anyone that wants to see it. But if you want to make a closed source program for linux its not a bad thing at all. But its normal for a small program to provide the source. You can even sell programs that are used for linux and no provide the source either. But for linux its normal to provide source and with windows its not normal.
The idea of linux is to provide a very good open source OS to anyone that wants to use it and provide that source for free. That cought on with the programmers so they will provide the program and source to anyone that wants to see it. But if you want to make a closed source program for linux its not a bad thing at all. But its normal for a small program to provide the source. You can even sell programs that are used for linux and no provide the source either. But for linux its normal to provide source and with windows its not normal.
Interested in being apart of a team of people that are developing a toolkit that can help anyone product an online game? Then click here http://tangle.thomson.id.au/
I guess it''s just my bad experience with binaries that led to that. From what I read, I must download a binary specific to my distro and processor type. And of course you have the library dependacies. That may just be a problem with the redhat .rpm installer style. I guess if they included everything they need inside the program and had no dependacies with any other program it should work fine...
Ah well, I guess game companies just haven''t caught on yet (being that most of them write games in DirectX).
I was surprised to see a game on the shelves at a local store that listed the OS requirements as Linux or Window the other day, so I guess linux/unix will eventuially have a decent library of games as more companies "test the water" with linux..
Until that day comes, I''ll just keep running linux inside windows with VMWare .
Ah well, I guess game companies just haven''t caught on yet (being that most of them write games in DirectX).
I was surprised to see a game on the shelves at a local store that listed the OS requirements as Linux or Window the other day, so I guess linux/unix will eventuially have a decent library of games as more companies "test the water" with linux..
Until that day comes, I''ll just keep running linux inside windows with VMWare .
"Mommy, where do microprocessors come from?"
RPM is a nightmare, but there are better solutions for packaging... for example, Debian and Gentoo has a package system that handles dependecies for you.
Victor.
Victor.
c[_]~~
they could also ship a set of binaries with an install script that determines what dependancies are not met and install them (i.e. have the dependancies on the disc)
further, i think the idea of releasing a game open source is alot like the idea of a shareware game (like say, doom originally was)
you release the binary opensource with a coupla levels of content, and then you sell on cd the installer, with all the dependancies nicely met, and all the rest of the game content
basically, you would pay for content, not the game itself
further, i think the idea of releasing a game open source is alot like the idea of a shareware game (like say, doom originally was)
you release the binary opensource with a coupla levels of content, and then you sell on cd the installer, with all the dependancies nicely met, and all the rest of the game content
basically, you would pay for content, not the game itself
-PoesRaven
Most games for Linux(many of which came from the now-extinct Loki) utilized their own installer. Quake3, UT2003, SoldierOfFortuen2, SimCity3000, Rune, etc.. all have their own custom install program. That install program was built by Loki and is open-source I believe. So if you have a game you want to distribute - you don''t have to use RPM just use the Loki Installer. It is very similar to the InstallShield installers you use on Windows and can be used for closed-source software.
Just an idea building on this: When you''re releasing a binary package for Linux, you could provide (meta)packages for distributions handling dependencies (like Debian). This way, the user would install an empty package that just has dependencies to the libs needed for your program, like SDL and other stuff so that updating the system or removing packages would be unlikely to break your program.
That way, you could add a function to your installer that tries to determine your distribution (or perhaps just ask the user) and then install that package. This would automatically lead to the installation of your dependencies.
Machine/library dependency could likely be handled by a web installer for the relevant binaries only. Pure, architecture-independent data and a basic set of binaries for the most common version(s) of glibc, libstdc++ et al. would be distributed on CD but you could add other versions as needed (i.e. when RedHat decides to introduce incompatible versions of the C++ library).
All said, the version multiplicity still sucks for binary-only distribution. The only potential fix-all is static linking.
That way, you could add a function to your installer that tries to determine your distribution (or perhaps just ask the user) and then install that package. This would automatically lead to the installation of your dependencies.
Machine/library dependency could likely be handled by a web installer for the relevant binaries only. Pure, architecture-independent data and a basic set of binaries for the most common version(s) of glibc, libstdc++ et al. would be distributed on CD but you could add other versions as needed (i.e. when RedHat decides to introduce incompatible versions of the C++ library).
All said, the version multiplicity still sucks for binary-only distribution. The only potential fix-all is static linking.
My experience with open source is that unless everyone releases their stuff open source, it doesn''t work like it''s supposed to, and the small percentage that release their stuff open source become the sacrificial lambs.
If everyone released everything open source, then everyone would learn new things and get to see how other people went about solving the same problem, and that usually results in significant learning and improvement. This is true in anything, not just software development. It is no coincidence that every chess world champion hung around with and studied with some of the other greatest chess minds in the world. Seeing how someone else solves a problem can completely revolutionize how you think about solving problems.
If only a few do it, then those few (if they''re any good, mind you) usually get left behind, or lose their edge at the least, because then the above phenomenon of revolutionary improvements happen...for everyone else. An example of this is the open source chess program Crafty. It was written by Dr. Robert Hyatt, who also wrote the chess program Cray Blitz, which won the computer chess world championship several times in the 1980''s. Crafty was one of the strongest programs available in the mid to late 1990''s, and now it is still among the top programs, but there are numerous other amateur programs that are equal or better in playing strength. If Crafty had been closed source, I wonder if all of these other programs would be as good as they are (I doubt it). However, Dr. Hyatt is a professor and probably cares more about the advancement of computer chess and artificial intelligence in general than his own success.
Releasing something open source is good for the overall advancement of a particular field of study, but it is not good for parties in a competitive environment, unless everyone does it (and even then it doesn''t make sense for the top party).
If everyone released everything open source, then everyone would learn new things and get to see how other people went about solving the same problem, and that usually results in significant learning and improvement. This is true in anything, not just software development. It is no coincidence that every chess world champion hung around with and studied with some of the other greatest chess minds in the world. Seeing how someone else solves a problem can completely revolutionize how you think about solving problems.
If only a few do it, then those few (if they''re any good, mind you) usually get left behind, or lose their edge at the least, because then the above phenomenon of revolutionary improvements happen...for everyone else. An example of this is the open source chess program Crafty. It was written by Dr. Robert Hyatt, who also wrote the chess program Cray Blitz, which won the computer chess world championship several times in the 1980''s. Crafty was one of the strongest programs available in the mid to late 1990''s, and now it is still among the top programs, but there are numerous other amateur programs that are equal or better in playing strength. If Crafty had been closed source, I wonder if all of these other programs would be as good as they are (I doubt it). However, Dr. Hyatt is a professor and probably cares more about the advancement of computer chess and artificial intelligence in general than his own success.
Releasing something open source is good for the overall advancement of a particular field of study, but it is not good for parties in a competitive environment, unless everyone does it (and even then it doesn''t make sense for the top party).
My game is closed source, and has no installer (just unzip, chmod, and run).
So far, no one had a single problem with (well, one did, but he didn''t read the faq.txt where it said that you need to have sdl_mixer installed (among other things)).
Height Map Editor | Eternal Lands | Fast User Directory
So far, no one had a single problem with (well, one did, but he didn''t read the faq.txt where it said that you need to have sdl_mixer installed (among other things)).
Height Map Editor | Eternal Lands | Fast User Directory
This topic is closed to new replies.
Advertisement
Popular Topics
Advertisement