Dan,
Here's some of the anti-Rambus hyperbole (positioning?) from your post that I find as grating as you all find the pro-Rambus hype:
Shipments at levels that are more than samples should be in the hands of end users starting some time next month. That beats Q1 of 2000, but just barely. It's been a long 3 years for PC rambus.
Since Intel chose RDRAM over DDR it's been an equally long 3 years for DDR. Probably longer since they've been fighting an uphill battle the whole way, whereas Rambus has been gaining design win after design win along the way.
VIA is showing pre-production silicon for DDR (which may just be a licensed version of Micron's samurai chipset - so I won't count Micron as another vendor) but AMD is also showing its pre-production DDR silicon.
Whoop-de-doo. Intel is showing production silicon for RDRAM. What's the big deal showing pre-production silicon? As AMD has shown for years, and as Intel has shown recently, there's still often a long gap between pre-production and production runs. Are you saying that the DDR chipset production from here is going to go without any hitches whatsoever? If so, then you're guilty of the same hubris that many of suffered with respect to the 820.
And of course then they have to get Dell, Compaq, HP, et. al. to actually use it. More hubris if you think they're just going to be standing in line to put DDR chipsets in their boxes. It's as much a marketing game as a technical game. Where's the driving force behind driving DDR acceptance? Who's verifying that all of the DDR DIMMS will interoperate? Lot's of work to do still. The "if you build it, they will come" mentality has led many companies into bankruptcy.
Good luck with such pie-in-the-sky thinking.
Just keep in mind that so far Semico has been right, and Dataquest has been wrong
You made this statement with respect to Semico's forecast that RDRAM in 1999 would have 1.7% of the market versus Dataquest's 5% forecast. Since neither company expected the 820 to be delayed and both forecasts were made when September was the expected ship date, this is a very self-serving statement. Semico didn't say that it would be 1.7% because there would be a delay -- they said it would be 1.7% because they didn't think it would ramp very rapidly. Since we still haven't seen any shipments, we have no idea how fast the ramp is going to go and you can't make any statements about either forecast. Certainly, since it appears we're not really shipping until December, it is extremely likely that the share for the year will be virtually 0%. But as I mentioned in the post to Carl, the thing about forecasting is to verify that, even if you're correct, you were correct for the right assumptions. And so far we can't test either set of assumptions. So at this time no one is right or wrong yet.
Dave |
