Slashdot Mirror
SCO Releases Linux OS for Itanium 2
GreyPoopon writes "Computerworld has an article referring to SCO's announcement of Enterprise Linux for the Itanium 2. Base installation starts at $999 for up to four CPUs. My favorite quote: "With its new system, SCO is a little late to the Linux on Itanium 2 market." I would think being late would be the least of their worries right now. I personally consider this to be my daily dose of comedy. Newsfactor has a better article."
More information on SCO, it's lawsuit, the many boycotts against it and why they aren't exactly the ideal company can be found here.
"I filter at +6, and have yet to miss out on an important comment." (#822545)
Perhaps. However, people also buy such systems for the memory bandwidth. Infact, I would expect that to be more likely to be the case than there being any real interest in exploiting the larger register size.
A Pirate and a Puritan look the same on a balance sheet.
Umm, last I checked, a 64-bit processor doesn't give you a wider bus automatically. That's purely a function of the attached bus itself. In fact, in the RDRAM architecture, the bus is actually really *narrow*, it just runs at very high clock rates (IIRC, in the Ghz range these days, with packets that are 8 bytes wide). So, no bandwidth is probably not a key factor in the decision to buy an Itanium-based machine.
The odds are good they either want 1) really excellent floating point performance (something I hear the Itanium is good at), 2) better performing native apps, because, theoretically, the compiler can optimize better with more registers and the ability to statically schedule multiple operations simultaneously (although we'll see how that plays out in practice), or 3) the ability to access large amounts of RAM.
Note, the second point I listed is a bit of a red herring, since I'm not aware of any really good IA-64 compilers yet (other than maybe Intel's), and I'm sure the practice of optimizing for the IA-64 is still developing.
Patents aren't bad. Neither are trade secrets. There are 2 issues that most OSS developers take w/ SCO's lawsuit vs. IBM.
1.) SCO claimed in their brief that Linux is an immature operating system, that could only have gained dominance in the marketplace by IBM stealing IP from SCO and putting it into Linux (which is just patently false... pun intended); and
2.) SCO claimed in their brief that OSS developers don't have the intellectual wherewithal to have developed similar features found in SCO's UNIX on their own, short of IBM lifting SCO's IP and showing the OSS developers how to do it, which is also patently false.
Additionally, proving that SCO IP was somehow misappropriate is going to be just short of completely impossible, again making their lawsuit vs. IBM appear more as a grab for either 1.) as much $$$ as a trial will award them, or 2.) a plea to be absorbed/taken over by a profitable company, such as IBM.
In this light, the OSS community doesn't necessarily take issue w/ IP per se, just SCO's flagrant remarks against OSS developers in their plea to become profitable or purchased through litigation.
IANAL, but if I recall correctly, the SCO vs. IBM issue is about trade secrets and contract violations, not about copyright or patents. Therefore the GPL clauses about patents and copyright are irrelevant in this case.