OS UPDATE - Part 2


by Wayne M. Krakau - Chicago Computer Guide, September 1998

Here’s the second in my series on the latest operating systems issues. This article covers the upcoming release of Windows NT Version 5.
As with Windows 98, I have great hopes of NT 5 fixing a lot of problems and omissions in its predecessor, though the problems I reported last month in Win 98 have me running scared.
In a manner similar to what was done with Win 98, some braniac put a notice in one of the technical areas of Microsoft’s Web site that NT 5 would fix more than 10,000 outstanding bugs in NT 4! (Yikes! Is there something in the water supply at the Microsoft corporate campus that drives people to commit ritual employment suicide?) Microsoft management immediately had this notice modified to have a more marketing-friendly phrasing - well, almost immediately - not until the national press found and published it. Oops! Again, no mention was made of the ethics of having released a product with that many bugs, or of the original bug count, prior to the release of the various patches and fixes that were subsequently released.
In this case, it was a technical person, not a marketing or sales person who released the embarrassing data, so he or she might not be demoted to box-stuffing duties, as a technical person would not reasonably be expected to be educated on the benefits of the Microsoft Creative Truth (in English: lying) policy. However, I suspect that the techie who did it might never again be allowed to even browse the Web, much less alter a portion of Microsoft’s own site.
Another thing that worries me is that the commercial-release version of Windows NT Version 5 is a ghost, an ethereal spirit, a figment of Bill Gates imagination - in other words, vaporware. This is not a piece of software that is within a few weeks of release, with all of the features pretty much locked in, and only the last minute removal of test code and the addition of the last few bug fixes pending. This is software with an expected (as opposed to officially announced) release over a year from now (as I write this in August of 1998).
Microsoft has incorrectly estimated the release date so many times that these inaccurate announcements now seem less like overconfidence or even incompetence and more like a specific marketing policy of what was termed in IBM’s glory days as FUD - Fear, Uncertainty, and Doubt. A major portion of this policy is to keep people from buying your competition’s product by dangling the imminent release of your allegedly killer product in front of them, even though it’s nowhere near ready for release. The idea is to either get them to buy your current not-so-great product in the hopes of an easy or inexpensive transition to the new one, or, to make them so worried about making a decision that they hold off buying altogether. If imitation is the sincerest form of flattery, then IBM must be very proud. (Of course, you need to remember that this aspect of IBM’s FUD policy was a factor in getting antitrust investigators to go after IBM years ago!)
Any product as far away from its release date as NT 5 is, could easily change by the time it is released. A particular feature that sounds appealing now might end up not working as advertised, be buggy to the point of being unusable, or simply be omitted from the final product at the whim of the developer. The last-second addition of other, as yet unnamed features, (especially in a company driven by marketing needs as opposed to technology or customer service) might also destabilize or just slow down the base product. You’ll have to consult your favorite psychic for the true facts.
From what I’ve seen, Microsoft is not going to change one important characteristic of NT that has always been one of my pet peeves. That is the use of a desktop operating system as a server. I still lean strongly in favor of true honest-to-God network operating systems that were designed and optimized as such as opposed to a jack of all trades (and potentially master of none) operating system that must be both a desktop and network operating system. (Or as a fake commercial on Saturday Night Live once put it, "It’s a dessert topping AND a floor wax, too!") I would be much happier to see Microsoft split NT into two separate products with two separate development teams, each with its own unique goals. A while back, Microsoft claimed that they were different, but some serious techies (with more expertise in such matters than I have) exposed that lie by examining both products with specialized software tools and finding that their kernels (the base underlying programs) were identical. Oops!
Microsoft is finally addressing the performance deficiencies of NT, but I am worried that there are limits to what can be done while still keeping this dual mode network/desktop policy. With NT 4, all of the information I have been able to gather from software companies, very large resellers (who have dealt with both NT and NetWare for some time), and system administrators indicates that the performance ratio between NetWare and NT 4 for file servers is 4 to 1, and for application servers (database, mail, etc.) is 6 to1. That is, if you have a maxed-out NetWare 4.11 file server and replace NetWare with NT 4, you will need to buy three more equivalent servers to handle the load (hence the 4 to 1 ratio). If you have a maxed-out NetWare 4.11 application server and replace NetWare with NT 4, you will need five additional servers (for a 6 to 1 ratio). I have seen similar performance ratios in the national press though with some variation in the actual numbers (3 to 1 for file servers in one article, for example). Obviously, anything Microsoft can do to reduce these ratios will help.
Interestingly, these ratios are part of the motivation for many NetWare resellers (as evidenced both by my own personal interviews with them as well as many nationally published reports) to jump on the NT Server bandwagon in spite of believing that NetWare is a far superior solution. They like the profit in selling more servers to handle a given need. They also like the additional profit on the labor side of the equation, too. They tell me that the additional time needed for installation and integration leads to some additional billable hours, but that the additional ongoing maintenance and management costs of NT versus NetWare are where the real profits are.
Next month, I will continue with my NT 5 coverage. In the meantime, please note that while I am a rabid NetWare fan (feel free to take my opinion with many large grains of salt), I know full well that market demands will require me to actively sell NT Server Version 5 (and maybe even NT 4) even though I only currently sell NT Workstation Version 4. I have multiple employees undergoing certification training in NT 4 now, and have the materials (though, at the moment, not the time) to get certified myself. While I have strong technical and ethical opinions, my MBA specialty was in Marketing with a sub-specialty in Marketing Research. Just as I am authorized to sell multiple brands of computers in order to satisfy my clients’ demands, even though I have an overwhelming favorite brand, I will need to satisfy my clients’ needs for alternative network operating systems. Besides, every NT 4 server that I’ve encountered so far has been configured just as poorly as almost all of the NetWare servers that I see. Opportunity knocks!

�1998, Wayne M. Krakau