That's because the man in charge of SCEI, Kutaragi Ken, is an electronic engineer by trade with no proper training in computer science. While he knows how to maximize any given circuitry, he doesn't give a shit about the programming aspect of the architectures he creates because he doesn't know any better. And people who "knows better" cannot criticize him because of his position and ranking within Sony.
Sony put together a crap architecture and survived in this generation, but I don't know how long Sony's luck will continue to hold.
no proper training in computer science ?
Deadmeat, Deadmeat...
First you still assume he basically single-handendly designed both HW and OS and API for PS2, PSX and PS3 ( still how come PSX turned out easy to code out of the box [sort of
] ? Kutaragi was the MAN behind PSX too it appears
), second you make a claim on his reputation and knowledge...
You can list all the exams he took in his university life and all the experience he might have accumulated in all these years ( he is not 20-25
) ? Wow, that is amazing my friend...
Now you say "he knows how to maximize any circuitry"... well but you also said that the whole PS2 HW ( both EE and GS ) is architecturally flawed and arguably a good design over-all in terms of elegance ( you called it a piece of garbage in the past )...
Also, going by your logic, Shigeru Miyamoto is to be held responsible of every small graphical glitch/flaw present in SM64...
A crap architecture... BEEEP! the major problem for developers was Documentation ( or lack there of, at least at the beginning for american developers ) and lack of efficient high level libraries/API and developer were presented with a quite flexible system with new features and new ways of approaching real-time 3D graphics with "not excellent" documentation and having to write to the metal ( or obtain bad efficiency ) due to the lack of good high level libraries...
The architecture makes much more sense than the Saturn architecture and ideally is a direct extension of the PSX HW which people loved...
PSX:
MIPS core + GTE ( fixed-point Vector Unit ) + MJPEG decompressor...
PS2:
MIPS core + VU0 ( VU also available as co-processor ) + VU1 ( stand-alone Vector Unit ) + IPU ( MPEG2 decoder )...
The do not sound that different HW wise...
If I gave you a nice ARM7 CPU ( like the GBA CPU ) and I told you "now build me an OS with Windows like complexity, GUI, multi-tasking, etc..." and I gave you a compiler with BAREBONE libraries... and I required you to write good chunks in ASM... even with the ease of use of ARM ASM, you would still have a problem ( btw, I'd give you also a crappy User Manual for the ARM7 CPU )...
Now let's change the CPU with something more complex architecturally...
see the point of my "dramatization" ?
Again, you are declairing PS3 doomed/piece of crap 2 years before it hits the market, you are basically calling Toshiba and IBM's as not having any balls and saying in the matter... and you're saying that even if we find much more times other people's names in SCE's R&D papers and patents, it all down to what Kutaragi knows how to do and that he is too dumb to even have a large team work and put in practice several of the ideas he has ( assuming your model )...
I am sure that David Kirk is solely responsible for GeForce FX not crushing the Radeon 9700 PRO and mr. Meyer is the sole responsible for any of AMD mistakes...
PS3 will not require developers to code to the metal out of the box with crappy documentation, the developers feedback has been clear enough I hope in that regard... I think we will move back to what we had with PSX in temrs of having to use HLSL and high level APIs, you are just oing to have a much more flexible and powerfule HW...
Are you really telling me that Sony, IBM ( which is investing billions of $ in internal Linux related R&D and Linux is rumored to be used [modified, customized] in Cell OS ) and Toshiba are not capable of providing an HLSL ( Sony has been investing for quite some time in the Stanford Shading Language ), a good OS and API support ( completed with nice compiler support ) ?
And if they provide that ( very "likely"
) you would still say that it would not be enough and people would be "abbandoned" to such a crazy and obtuse architecture ( which is not ) ?