Hi Dave B; Thanks for the compliment. Now that we know that rdramreview is a secret Rambus employee/ insider/ exemployee website, (whether the owner is still employed or not), it makes it a very attractive place for me to work over, and their patent information is fascinating.
The whole site is full of information, it will take a long time to go through it. This will take weeks, and the arguments afterwards will, no doubt, take months. I hope that everybody can stand the wait. I am sorry that people lost money today, and I am sorry that I cannot write a complete analysis of that whole website in just one night. But you guys will just have to wait. (By the way, I am trying to get back onto a sort of normal schedule, so I will probably end up posting more in the early evening instead of the graveyard hours. The 7-11 where I work keeps moving my schedule around. :)
Eventually I will get to the patents, but they take a lot of time to analyze. You have to sit around at the IBM patent search engine and it just eats up a lot of time. I also have to go through reams of obsolete data sheets. In addition, I have had some difficulty finding data sheets for memories from the late '80s. Texas Instruments seems to be best for this. The reason I need those data sheets is because of prior art arguments.
In the process of talking about these patents we will have to cover exactly what the difference between an SDRAM and a DRAM is, and what consequences that has on the control of the DRAM. It's a very big topic. To put it in perspective, RMBS is reporting lawsuit costs of high six figures per quarter.(?) Even at my wage rates, that is a lot of hours, though at least I don't have to file court papers or wear a suit. (I'd probably spill a slurpee on it, anyway.)
It is now clear that the DramReview site is anything but unbiased. This should give people who have been assuming that their calculations are correct some pause. I would like to go through the whole site, but will start by showing that their power calculations are a lie because these are the easiest to calculate. They give a set of patents that show very good coverage of SDRAM techniques, I think, but if the patents shown were selected in a way at all similar to the way the DRAM power consumption was calculated, then there are probably a lot of unmentioned patents. The fact is that the US patent examiners don't know crap about memory. We will have to do long searches for other patents, and for prior art before we can look at the patent situation.
-- Carl |
