Private lives

In his "Backspin" column about Intel's serialized processor chips ("Jump on the serialized bandwagon," Feb. 1, page 58), Mark Gibbs says, in essence, "Never fear, lots of other computers, such as mainframes and minis, already have readable serial numbers. It's no big deal, really!" However, mainframes and minis aren't personal devices. Lots of people use shared servers, so information collected about those servers doesn't shed light on the behavior of an individual. But once programmers get their hands on the unique serial numbers branded into our personal desktops, those serial numbers will begin appearing in all sorts of databases, letting people collect and exchange profiles about personal computing habits.

Gibbs says a serialized CPU is no more dangerous than the Ethernet address on a PC. But the Ethernet address is alterable on many network interface cards, and in most organizations Ethernet cards hop from machine to machine like fleas on dogs. Nobody uses the Ethernet address as a tracking ID because it's not reliable. Intel says its CPU serial number is immutable, which is just what the snoopy people are lusting for.

I don't think the Electronic Privacy Information Center is being at all premature in calling for a boycott of Intel products. If we don't stop Intel from serializing CPUs, we can all kiss our desktop privacy goodbye.

Mark Gibbs claims we have a long way to go before processor serial numbers become like Social Security numbers. But that seems to be how Intel is advertising processor serialization. To me it sounds like Intel is planning to broadcast these serial numbers every time you visit a Web site. That makes processor serialization more powerful and more dangerous than my network interface card (NIC) address or TCP/IP address. And if my NIC address is unique, what do I need a processor serial number for, anyway?

Has Intel announced any productive use for this serial number? The only uses I have heard about so far are Web tracking and copyright protection. Are there any uses that actually benefit the user?

Security is almost a moot point. It's just another number that would be added to an encryption scheme. And if Intel lets Web sites capture the number, how is that any more secure than what we already have?

Sounds to me like Intel doesn't yet know what it wants to do with this thing. All I know is if it has no clearly defined purpose or value to the end user, I don't want it.

Your article "Novell Reborn" (Feb. 8, page 1) was an excellent piece of industry journalism: well researched and comprehensive. It's great to see Novell getting positive front-page press on Novell Directory Services and its rebounding market share. I hope more and more people come to realize that what Microsoft does best is marketing and not software development.

Just wanted to compliment you on the thoroughness of your article on the rebirth of Novell; it was definitely well-written and informative. Believe it or not, there are still a lot of shops out there that are using multiple platforms and somehow managing them all with limited staff. Anything positive on Novell is more ammunition I can use to convince management to stick with our current Novell servers and upgrade rather than migrate to NT.

I prefer the user-friendly Novell Directory Services structure; however, I haven't been given the opportunity to try Novell NetWare for NT on an existing NT server.