Tim notes: Those boot times are from the same machine - a celery 566 overclocked to 707 with 256M. The 36 sec boot time is after the BIOS post and the OS starts loading. It ends at the functional desktop. In this time, it loads an IRC session, ZoneAlarm Pro, McAfee VirusScan, a FTP Server, Apache and I'm sure I'm missing a few things.

Yikes, that's fast. I *am* jealous. Even adjusting for slower machine (with 128 RAM) and ~15 seconds of POST, this Win2K isn't even close. Not there to check, but I don't think my Win2K home PC (866 P3/512 RAM) is much faster on boot than this laptop. Boot times seemed like less of an issue with Win2K power management and suspend capabilities, but that's one of the things that isn't working as well as it did at first.

(.....) As for benchmarks, when Win2k first came out, there were benchmarks that showed that NT4 was actually faster than Win2k when both had memory over 256M (or maybe it was 128M). Not many people seemed to notice, or care. I don't put a lot of stock in benchmarks, whether on accident or with malicious intent, benchmarks can be made to show whatever you want.

Like any research study, I think these come down to whether you are measuring the right things, whether the methodology is made clear, and whether the methodology is legitimate for what you're trying to measure. Absolutely agree that interested parties can structure these to show what they want -- witness the endless go-round of "fastest" RDBMSes --, but I also believe that it is possible to perform benchmarks that are useful. Infoworld and their ilk aren't exactly peer-reviewed, though, and the methods aren't always spelled out as well as they should be. Even if I do have the time to critically read the whole thing, I'll tend to take each one of these as a single data point on a scatter plot. That Infoworld *seemed* pretty reasonable, but it was focused on a small set of multitasking benchmarks. I'll keep reading.

I *have* spent a bigger chunk of my life than I would prefer responding to execs who throw a benchmark in my inbox and, without having read it (Understandable. It's not their bailiwick.), demand that I debunk the headline that usually says something like "The Product We Are Using (The one Jim recommended!) Sucks!"