Lol,

I realized that we''ve covered so much stuff we probably skipped Dungeon Master''s questions =)

I think the general consensus is that the majority of the game market is Console or Win32 based. A lot of companies get some pressure and demand for OS X and Linux game versions, but they make the bulk of their money on Win32.

Doesn''t mean there isn''t room for games for Linux, actually quite a lot of buzz from people who want to see them for their Linux desktops. I also seem more and more companies using Linux behind the scenes of online games versus MS servers. It also seems to me that more and more PC game makers are promising Linux and Mac OS X version to their customers. They don''t always deliver of course, but they seem to be under some pressure which indicates their is a demand or growing demand.

I think you can probably also pick technologies that are a little platform neutral, or cross-platform. OpenGL vs Direct3D, SDL, OpenAL, etc.

Interim

Is there a personal hobbist reason to have linux just purely for game programming, not to market your games just for fun!

Do you Personaly make games on the linux O.S. or because its not Microsoft and something different than what is widely used.

Having windows to make games because of directx and I know that you can use OpenGL on linux but I like to have both.

On another note:

I agree that Desktop computers are disapparing and replaced with intergrated items. Because of abundant technology built into everything to you cars GPS system, Computerized refrigerators, PDAs, Cell phones which are mini laptops.

The race of the bulky desktops is endangered by its own technology.

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Ignore me if I sound Dumb!

quote:

---

Original post by CmndrM
The Bad:

...

X. (I just plain don''t like it. It''s slow. However, starting with the 4.0 release, they have imporved it greatly.)

---

Yeah, X is a problem. For me, though, the problem is that its architecture has failed to evolve with the times. For example, its network protocol is too "heavy" because it needs to send the information to describe every line, every color, every pixel. A smarter protocol (which could be "inserted" between X client and server to reduce bandwidth requirements) would first describe high-level objects and then transmit events that altered the states of these objects. The other reason I advocate this approach is that it serves as a perfectly compatible transition period to a more compact, lightweight GUI engine.

quote:

---

Original post by Interim
I think the driver issue isn''t as bad as most Linux opponents point out. It''s really just manufacturers doing numbers, "I got 98% on Win32, 2% on Linux. We got 2 months to market, just do the Win32 drivers." I know this happens with my company with Linux projects we push for. It comes down to time and what is the best way to hit as many customers as once. Linux just doesn''t have the market numbers yet to drive the business cases.

---

True as that is, the problem is with the manufacturer perception that all work must be done by their engineers. Linux developers will happily write you a device driver (a high quality one at that, thanks to open peer review) which you can bundle with your hardware. Some will even agree to do it closed source (though not many). The problem is that manufacturers feel that the kinds of specifications necessary to write a good driver yield proprietary information, which they don''t want falling into the hands of competitors. If they can find an arrangement where the developers sign an NDA, implement the driver and keep the source closed for a stipulated period (for example until the next product iteration, or until the product is superceded by a better one), then we may completely solve the driver availability issue even without profit viability.

quote:

---

Problem with the organic vs unorganized is the lack of application interfaces. Linux/Unix is great on the CLI in this regard. It had a simple philosophy of specific tools chaining together. I don''t know if it''s the complexity of the GUI, but it seems to fall apart there. Mac is great in this regard, Windows is ok. Mac incorporates things like Spell check into just about every application since it''s available as a "freebie" in their application framework objects.

---

A few friends and I have been mulling the overall software development model, where each application is responsible for implementing or integrating the code it requires to provide certain functionality. From a binary perspective, how many copies or implementations of, say, a GIF reader might the average user have on his/her system, irrespective of OS? With a loose-coupling, late-binding system of reusable OS components, any application can take advantage of any functionality that is installed - and any installed functionality becomes available to any application.

Let me take a moment to explain this further. If we categorize data based on type, using a MIME-like hierarchical organization of data_type/data_format, then an application deals generically with data_type while format-specific plugins convert from data_format to generic data_type. So, if IE for NewOS can render images and J. Random Hacker creates a new image format, JRH, then for IE for NewOS to be able to render JRHs all that needs to be done is for JRH.NewOS.plugin to be installed and all applications that process images (fortuitously including IE for NewOS) can read and write JRHs. This model moves application developers to dealing with application functionality - real functionality as opposed to I/O capabilities (ever see those older software ads that say "reads over 300 file formats!"?)

Okay, so we apply this philosophy to our GUI subsystem and whenever a tool/component/plugin is installed, it becomes available to all applications. Every text editing application can then use the installed spellchecker, even though the spellchecker isn''t an intrinsic system component. And so forth.

quote:

---

I do see Linux being a place in this trend, but not as a desktop, but as the OS on your PDA, your phone, or as the OS running your toaster and microwave. It''s organic, compact nature is excellent.

---

Yes, I used to advocate the invisible OS philosophy as well, but I''ve been wondering about that recently. There really is no reason why every GUI on every system should be identical or similar - externally consistent. Really, all we need is for the system to be consistent internally - for each user to be able to interact with all apps on the local machine in the same way. For that kind of idiosyncratic flexibility, there''s obviously only one choice (on a social level; as I said before, the BSDs fulfill this role admirably but lack the community wackiness to make them automatic "shoe-ins"). Linux on the desktop will emerge when a lot of today''s assumptions about the nature of desktop interfaces are genuinely questioned. It''s like the discovery of public-key encryption was tied to questioning the de facto assumption that the encoding keys must preexist the exchange and already be in possession of the parties wishing to communicate successfully.

Put succinctly, when we ask both "why" and "why not", we find new things.

The current "personal computing" paradigm is another thing we need to question. Computers are permeating our lives, much like telephones and cable TV have before them. How much longer until we start building homes with a view to computing? How do we do that, you ask? Well, when a home has a server installed in the basement acting as a data repository, with your house key containing digital keys that encrypt your data (this would be a key you create rather than some key given to you by a landlord), then you can fully distribute your interactions with the system across appliances like cell phones, PDAs, consoles, TVs, stereo components and so forth. The only remaining vestiges of our current PC model will then be the keyboard and some form of rapid pointing device.

Of course, right now I''m sounding more like a science fiction author (the lame kind who focuses more on tech than the effects of that tech on society), so I''ll cease my conjecture and move on.

Linux already leads the IT closet, even in organizations that are staunch Windows users. I read a letter in Linux Journal (and I don''t usually give much credence or attention to anecdotal evidence, but this one was funny) about an engineer who hacked up a tool on a Linux box for his company that performed fairly mission critical tasks on the ultra-cheap. After a bit of resistance, the machine was placed in the company server closet and everyone went back to work reaping the benefits of this machine. Then one day it suddenly went down. Only after much abuse for the unreliability of his "toy" was it admitted that the system was restarted because "well, it had been sitting there for a couple of months and no one touched it so we assumed no one was using it." In other words, these techs are used to literally babysitting their servers. Once prejudice is set aside, the reliability benefits alone make Linux (or the free BSDs) a no-brainer.

Finally, for up-to-date kernel info (because my summaries are really rough around the edges), check out KernelTraffic.

quote:

---

Original post by D-ungeon M-aster
Is there a personal hobbist reason to have linux just purely for game programming, not to market your games just for fun!

---

Sure. Linux is amenable to several platforms, so you could tinker with that old Dreamcast in your closet, for example, trying to get your game running on a console - or the XBox. You can''t do that with any other OS, currently (there''s a Dreamcast Open/Net/FreeBSD effort, but it doesn''t seem to have significant momentum).

quote:

---

Do you Personaly make games on the linux O.S. or because its not Microsoft and something different than what is widely used.

---

Personally, I don''t write games on Linux. I''ve thought about buying an old used Dreamcast and hacking it, but I''ve decided it''s not worth my time. The items I''ve mentioned here - evolving the GUI, involving the development paradigm - are more important and interesting to me right now than gaming. Plus, I think the future of gaming lies in the console because of the more natural form factor (couch, TV, friends as opposed to desk, PC, alone), but that''s just a personal opinion.

quote:

---

Original post by D-ungeon M-aster
But is linux good for Programming games?
or is it best to have both O.S. for options?

---

For games, you might as well do both.

Coding X straight is a pain. It hurts a lot*...
Usually, most people that I know, use some sort of well documented toolkit, such as mxWindows, GTK+, SDL, etc. If you know your going to use a kit, you might as well use a portable kit. And if you use a portable kit, you might as well port it to Windows.

Most Linux freeware games get more play than the average Windows freeware game. (I bet if Ztank or XBill was for windows only, no one would have heard of them). So even though you have the potential to reach a smaller audience, you have a better chance of reaching a larger one. Plus, since you have a windows verson, you don''t lose out on the larger audience.

On the engineering side, if you code so your program is portable, than you usually tend to write better code. It forces you to plan better/smarter.

At least, that is how I see it, your milage may vary

(*Not very well documented, the complexity is about the same as a Win32 application, but without the helpful MSDN, or as many tutorials)

quote:

---

Original post by Oluseyi
I think the future of gaming lies in the console because of the more natural form factor (couch, TV, friends as opposed to desk, PC, alone), but that''s just a personal opinion.

---

To go on a little tangent, I (mostly) agree. I used to think PCs were much better for games. But lately, I''ve been completely changing my views, not only because of the greater social aspect of console games, but also from a business perspective: It''s a lot easier to design a game for a piece of hardware that''s invariable than it is for a piece of hardware where the specs vary widely from the low-end to the extreme high-end.

On PCs, the designers must decide what type of machine they want to target. On the one end, they sacrifice graphics, sound, and general effects in order to allow more people to be able to play it. However, while doing this, the other companies that sacrifice the lower-end people to have as realistic effects as possible draw away the people with the high-end computers. On consoles, however, this problem is non-existent, since everyone with a Gamecube has the same piece of hardware (with the exception that one person may have a wireless controller and another would have the corded one. A very minor exception, obviously, since it doesn''t change the programming at all). A PS2 is a PS2 regardless of who has it.

The point of all this is that the social aspect is not the only reason why console games will (and already are starting to) dominate the market. The market itself also helps to determine this, as it''s much easier to design for one specific piece of hardware than it is to design for a whole plethora of options.

To be quite honest, the only areas of gaming I see PCs dominating in for the time being are online games (since console online gaming still lags behind a great deal), and certain genres of games that don''t lend themselves well to controllers (e.g., RTS games. There are a few other types, as well, however). Of course, the latter of these can be solved by creating new types of input hardware for consoles that can also become standards, and the former will eventually be solved as consoles are brought up to speed in online gaming.

quote:

---

Original post by Oluseyi
Let me take a moment to explain this further. If we categorize data based on type, using a MIME-like hierarchical organization of data_type/data_format, then an application deals generically with data_type while format-specific plugins convert from data_format to generic data_type. So, if IE for NewOS can render images and J. Random Hacker creates a new image format, JRH, then for IE for NewOS to be able to render JRHs all that needs to be done is for JRH.NewOS.plugin to be installed and all applications that process images (fortuitously including IE for NewOS) can read and write JRHs. This model moves application developers to dealing with application functionality - real functionality as opposed to I/O capabilities (ever see those older software ads that say "reads over 300 file formats!"?)

---

The Amiga had this and it was very cool. The implementation was a bit clumsy and it was catching on as the Amiga died, but as a proof of concept I think it was a success.


Stay Casual,

Ken
Drunken Hyena

good:
Customizability, it breathed new life into an old laptop when a custom compiled kernel droped the mem requirements to well under the laptops 48M ram. Win98 barely ran on it.

Control, I can control everything on the system. I''m not stuck only being able to do what MS wants me to be able to do.

Price, I''ve actually bought a couple of distrabutions over the years just to help put some money into it even thought I could have downloaded it for free. I don''t mind paying for software but $300 for each install of the os and $500 for each office package. On my 4 computers that would be $3200! I''m not that rich.

Bad:
lack of the good games. I don''t mind using wine but that still wont run most of them. I wanna play C&C Generals without rebooting to winblows.

X is too slow. The fact that it works through a network interface rather than direct commands like windows slows it down and that can sometimes be a problem.

The differing standards for configuration of the system. The fact that different distrobutions will install the files for the same program in two completely different places. I''ll admit that this has been disapearing recently though.

The part of the linux comunity that seams to think that anything that is even similar to anything for MS is crap. This is bad because I actually like how some of the software for MS works and most people aren''t gonna relearn how to use there computer.

Those people that think having to pay for software is a crime. The ones that think everyone should have to release there source code.

I know the last to will probably start somthing but it my opinion.

Oluseyi:
Sure. Linux is amenable to several platforms, so you could tinker with that old Dreamcast in your closet, for example, trying to get your game running on a console - or the XBox. You can''t do that with any other OS, currently (there''s a Dreamcast Open/Net/FreeBSD effort, but it doesn''t seem to have significant momentum).

D-ungeon M-aster:
I did the same thing to my old Dreamcast and X-Box!
but I used windows for my Dreamcast Emulation(winCE). I will try to use Linux on my Dreamcast.
I used Linux on my X-Box from the project on sourceforge.net
and I also use the PS2 Linux kit, its a bit pricey but worth it!

Added:
Linux: Changeable to your heart desires!
Windows: Good for beginners but old and stuffy for someone who wants to have fun!

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Ignore me if I sound Dumb!

I just installed RH9 and it is clear that the main linux distributior is not really interested in the desktop. The default settings is perhaps acceptable for servers but running makewhatis and similar every day is annoying on a desktop computer. OK, its easy to move some stuff from cron.daily but I suspect that there are others not so easy to change settings.
They could also distribute the drivers from NVidia or at least the auto downloader program. I dont know if they are uninterested, incompetent or just plain stupid.

Despite the above bashing is my opinion that linux is the best platform for development.

quote:

---

X is too slow. The fact that it works through a network interface rather than direct commands like windows slows it down and that can sometimes be a problem.

---

X is slow, but not for that reason. XFree86 has been rewritten so as not to use the network, but it was found that the loopback was definately not the bottleneck.

X is slow because of the drivers. Redhat, last I checked, doesn''t configure GLX or DRI for nvidia cards (It might not be possible without the kernel and GLX drivers from nvidia).