Soon. Its growth rate will begin to approximate that of PC's, which I believe is projected to be in the high teens, and may not even achieve that if Asian problems persist. The upgrade cycle will be broken as buyers conclude the incremental benefit to new versions isn't worth the cost (especially if Microsoft continues to raise prices, or if quality deteriorates further -- see below). Users will stick with the current release as "good enough." The upgrade model will be more like that for copiers, fax machines, file cabinets, and typewriters (after all, that's the kind of office equipment functionality PC's have replaced or augmented). NC's will also make a dent, as will alternatives like Solaris, OS/2, Rhapsody, Linux, etc. (Java will help make the choice of OS less determined by compatibility issues, making these alternatives more acceptable and better able to compete on technical merit and price.)
Once the stock drops and options as compensation become less available (the courts may have a role in this latter item by piercing the "temporary contract employee" fiction and requiring that all full time workers receive full benefits), MSFT will have trouble attracting employees who will work from dawn to dusk at low pay under the assumption that they'll get rich via the stock, and quality will drop still lower.
BTW, the recent 12/8/97 issue of LanTimes, at page 60, makes the interesting point that NT 5.0 increased in size over NT 4.0 by about 14 million lines of _new_ code in about 500 days with about 400 developers. I think that works out to on the order of one line per man-minute. The question the article raises is how complete quality control can have been.
Of course, that's just my humble opinion, and I know you and I don't always agree <g>. |
