cybamerc said:
The adoption of OpenGL isn't going to magically make the PS3 easy to develop for. Building on the work of others just gives Sony a place to start. PSP is a simpler design, uses an OpenGL like API, comes with many dev libraries... yet developers are still complaining because it's not working as intended.
Ehm - first, care to name the complaints of "not working as intended" or the devs who are complaining about them?
Secondly - if you actually believe that there ever was ANY early dev environment for any console, that did not generate any complaints, you're living very far detached from reality.
The only real issue with PSP development has been with the late arrival of final hardware - but that has nothing to do with development environment itself.
[quote="mrklaw']So what do people expect from PS3? We seem to know about Cell, but what graphics tech do you expect to be associated with it? Will Sony drop the ball by missing out things like pixel shaders, or other now fundamental features, instead happy to plough their own furrow?
And bearing in mind Sony seeming happy to ignore the PC graphics card power struggle and architecture paths, does anyone expect anything really left-field? A new way of doing things?[/quote]
At this point I no longer expect anything particularly left-field. Some more recent rumours do suggest possible surprises, but not really in the way most people expect.
Anyway - if Sony diverted from fragment shaders I wouldn't consider that dropping the ball in the least, there are tried and tested alternatives with known performance and other characteristics. Personally some of those might be entertaining to see implemented - I don't know if economical or wise, but just entertaining for seeing those less keen on adapting cry&complain about it.
But ok to summarize...
Pros:
- probably the highest amount of general purpose processing power.
- probably the most memory
- highest capacity optical-media
- probably default paper spec winner
- possibility of IBM development tools
Cons:
- still unanswered questions about programmability that could proove detrimental to potential performance
- depends on who is ultimately in charge of GPU design, but I worry about overengineering at expense of certain basics
- possibility of IBM development tools
I know this is all vague but I'm basing it on actual vague info out there, and I wanted to stay away from non-hw issues, which are more opinions then anything.
And no, the IBM thing isn't a typo - it could be both + and -, in a big way.