Actually, yes, I have 30 years experience with processors and systems. I started back in the 70's with the Zilog Z80 chip. As for the success of the cell, I will not argue with you. Look it up, the cell failed miserably. As an engineer I am very aware of optimization, it has been a large part of what I have done all my life. You however, seem to be confusing applications (game-GT5) code optimization with system (firmware) code optimization. Sony has no intention of updating the firmware or hardware for the PS3, especially with the new system coming so soon. That being the case, the aps that run on the system are extremely limited in what they can accomplish with code optimization. You are right, we will have to wait and see, but I wouldn't hold my breath on anything earth shattering with a new game on 7 year old hardware.
That's all fair enough, and I'm actually baffled by your over-zealous outburst regarding the Cell now, or maybe there's just some politics that I'm missing. Sure, Kaz isn't an electrical / electronic engineer with 30 years' experience in the field, but he doesn't have to be when his dev team can tell him what they're doing on various bits of the silicon, which
is their area of expertise. He has years of experience of listening to that information, and hearing the way it evolves during an ambitious software project and being able to compare that information with the actual results he sees in front of him. I'd suspect he knew exactly what he was saying.
As for the Cell being successful, I think that rather depends on what it was intended to actually achieve. The fact that it was different must be a success in itself, and although the PS4 has switched to good old x86 (from good old Power), it's technically got a great big array of vector units slapped onto the same die, too, which is going to be doing both what the SPUs did for the Cell and what a "classic" GPU does, only using a single instruction set that more coders are actually familiar with (although that wouldn't have been a problem had Sony actually had decent support in the early days of the PS3; they improved that dramatically, of course, and the devs reaped the benefits, as did we.)
The problem with calling the Cell a failure is that it's part of a much larger context that, overall (I used the phrase "Cell project" deliberately), was actually tremendously successful in opening up massively parallel computing in games for uses beyond just pushing pixels (something AMD has long been keen on, too). The chip design itself might have been a dead-end, in IBM's own words, but some of the concepts (like the parallelism and heterogeneous architecture - same as AMD's "Fusion") have been carried forward into newer designs - also IBM's words. Also, once developers got a handle on what that heterogeneous direction could achieve, they started asking for more of it, which is good news for gamers.
I won't confess to know the difference between firmware and software optimisation, and I'm not sure why it's relevant, as I was only ever talking about the game. The fact is that PD have had 2 and a bit years to work some more on PS3 (assuming they have been doing) and there's plenty they could have done in that time to make a better game, either technically, in terms of code or pixel shaders or whatever, or of the other stuff that goes into making a game.
As mentioned above, a modest technical improvement, like we've seen on previous hardware, is all that's needed, and all that should be expected. An improvement to workflows and the general artistic direction of the now-better-harnessed technical aspects of the game will result in a better looking, feeling and sounding game, too; but it's the game-design part, i.e. the part we actually play with, that's most important...
