[Open-graphics] Re: Open-graphics Digest, Vol 2, Issue 47

[Lars Thuring]

miller at techsource.com wrote:
> Kent Rosenkoetter wrote:
> 
> 
>>>The code morphing software really is SOFTWARE, but I've never seen
>>>people demand that it be opened up.  It's the big value-add that
>>>Transmeta has for their processors.  What's important is that they
>>>release specs for everything you need to know about how to USE the
>>>processors, even tips and tricks for optimizing your code to get better
>>>morphing results, etc.
>>
>>
>>I think the appropriate analogy is that we need to keep the ISA open,
>>not the circuit layout. No one badgers Intel for the internals of the
>>Pentium. Why should we badger nVidia for the internals of the GeForce?
>>All we need are the fully-documented interfaces. The ISA of the
>>graphics card. (And for programmable cards, the ISA of the shaders.)
>>This is what so many hardware vendors won't give up, and this is what
>>this project was originally predicated upon. Right?
> 
> 
> EXACTLY.
> 
> Someone reminded me that the original design had a programmable setup 
> engine (which was later dropped from the spec).  In the future, it will 
> reappear.  How to program that will be fully documented.  Doing that 
> sort of thing makes MY life a lot easier and gives developers a great 
> deal more freedom.  (It's amazing what people will do with your hardware 
> that you never conceived of.)

Exactly! ;-)

There must be a huge number of people intrested in a free FPGA design 
for doing 2D gfx which *never* gets close to being displayed using a PC 
gfx card. That is what attracted me atleast. The new page at 
http://www.opengraphics.org/ clears up quicker for everyone first 
locking at this (w/o reading 1MB of archives first). Still some 
intresting discussion here.

/Lars