Kash,
Thanks for your response and insight. You have generated some interesting and informative discussion about the future of the industry.
I would like to explain a little more what I meant in my post by "design methodologies" and why I think they are a key differentiating factor. I wasn't really referring to the design language or any of the EDA tools, although I realize that's probably what's more commonly meant by that phrase. So I apologize for being imprecise in my terminology.
I was talking about philosophies and practices of design. For example, do you use BIST, scan, Iddq, functional vectors, or some combination for testing? What is your clock distribution architecture? How do you adjust these things as the complexity of your designs grow? How do you coordinate multiple design groups working on the same chip across the country? What is your design debug process, and how many tapeouts do you have to make before you get to the final revision? Do you port your design to multiple fabs, and if so, how much of your design resources do you dedicate to this process? Are your various design groups sharing the insights and gotchas that they run into, or do you repeat the same mistakes over and over again throughout the company? Maybe I should have called it project management, although I was referring specifically to the design phase of the project and not to marketing, manufacturing, sales, or any of the other parts of the total project which have to be coordinated to make a successful product.
I have never been a designer myself, but I have been intimately associated with the process enough to see brilliantly conceived projects staffed with top-notch designers get derailed by the type of issues I listed above. If new product design is the heartbeat of the company, poor design practices are like arteriosclerosis. You can ignore it for a while, but it will eventually give you a heart attack, and if you don't reform your ways, it will kill you.
Regards,
G.P. |
