quote: Original post by Chai Peddler
Just as DirectX handles a variety of hardware configurations, the protocol would handle a variety of processors and OS''s.
At this point, I have to wonder what you mean. A protocol simply means some sort of standard or formal method. Some of these were already listed above... are you trying to say that instead there should be just one, for everything? Why would anyone compile their C++ to conform to your protocol when they could just compile it for a target platform and have it run faster?
quote:
This might (I imagine) require hardware compliance from the hardware manufacturers. Just as Direct3D support must be built in to 3D graphics cards to allow Direct3D to do its voodoo.
Not really... although undoubtedly cards these days are built with Direct3D in mind, they are not physically built with D3D support as such, that is largely a driver issue. The driver model follows a protocol...
quote:
As far as computer vendors seeking the cheapest OS... well, if enough support is garnered for this new protocol, they would be forced to include it, right?
Sure, but you''re avoiding all the real issues. Where would this support come from? Most PC gaming companies don''t care about non-Windows versions of their games as they don''t sell nearly as many copies. So they are unlikely to want to change their system. The OS manufacturers all cater to different audiences, and will not be interested in compromising to make some neutral grey mass of an OS. So they won''t be pushing it either. Most users don''t care what they are running providing it runs their software, so they won''t drive the revolution. Hardware manufacturers will also not necessarily make it easy... when a chip maker adds new instructions, if you can''t program directly in assembly you can''t take advantage of those.
I think that a combination of a standard programming language, and a cross-platform library is as close to a ''OS-independent protocol'' as is possible, given the state of computing in general.