FRAMINGHAM (01/31/2000) - In the network community, it is safer to talk about politics and religion than network operating system (NOS) preferences. Nothing brings out the vitriol more than a good debate about the relative merits of Novell NetWare vs. Microsoft Windows. If you don't believe me, log on to Network World Fusion and check out the tongue wagging in the forum that goes with our "King of the NOS hill" review, which ran last week.
Among a few kindly worded admonishments about not fully agreeing with our results, were entries like this:
"You gotta be kidding with these results."
"You must be on drugs."
"Too much bias - canceling my subscription."
And my favorite, which simply said: "Oh dear!"
Discounting the lunatic fringe and their conspiracy theories about Network World being controlled by Microsoft Corp., criticism seemed to fall into two broad camps.
One group seemed to have trouble digesting the fact that some of the 12 scores we awarded each NOS - NetWare 5.1, Windows 2000 Advanced Server, Red Hat Linux 6.1 and SCO UnixWare 7.1.1 - were based on benchmark tests, while the others were based on a subjective review of features supported.
Although we say in the story what was tested and what was not, in retrospect we could have made that clearer, perhaps by breaking the piece in two and labeling one section "test results" and the other "opinion."
The other major block of criticism took exception to the subjective scores we handed out for scalability, security and stability/fault tolerance.
Admittedly, "stability" was a poor choice of words on our part. The only way a system can earn that badge is a few years of hard labor in the real world. But more importantly, these results simply reflect the reviewer's opinions about product features which, in the case of fault tolerance, meant looking at things like RAID support and memory protection.
Did we get everything right? That debate is still raging online and, truth be told, that is part of the value of this kind of exercise. The discussion among experienced people adds context to the original research.
Curiously enough, lost in most of the online debate is the fact that Novell won the performance category. People don't seem to care, caught up as they are with the idea that someone found something to like in Microsoft's forthcoming product.