by Wayne M. Krakau - Chicago Computer Guide, May 1997

This is a continuation of my article following up on the subjects of past articles.
Migratory Verticals
I have encountered a trend lately which overlaps two of my previous subjects, the continued sales of NetWare 3.12 in the face of overwhelming evidence of its obsolescence, and the difficulties in getting good support from many vertical market software vendors. Companies are installing new 3.12 systems not because of poor advice from resellers (the most common reason), but because they are dependent on vertical market software that has not yet been tested and is not fully supported in a NetWare 4.x environment!
I can understand a software vendor’s reluctance to work with the old 4.0, 4.01, and 4.02 products. They were definitely "not ready for prime time." NetWare 4.1, however has been available for quite some time and is a proven, reliable product. Its descendant, IntraNetWare, is an even more capable derivative product. There is no excuse for neglecting this industry-standard product family for such a long time.
The client companies are naturally reluctant to become unsupported beta testers, so they buy decrepit old NetWare 3.12. Who can blame them? I can only hope that customer demand will eventually force these software vendors to update their products (if necessary) so they can test and approve them for use with NetWare 4.x.
The funny thing is that I found that the reluctance of some of these vendors to do this testing is based on horror stories about the original version of the workstation interface software, called VLMs (Virtual Loadable Modules), released with NetWare 4.0. VLMs long ago were stabilized, and are now so old that they are in the process of being replaced! If you are going to sell a NetWare-based program, at least you should keep up with the latest information.
ET Phone Home
After examining many computer telephony systems, and writing about several, I finally jumped on the bandwagon and got certified in CallWare (CallWare Technologies, Inc., 801-486-9922). After an intense week-long course, most of which was on computer telephony as a concept, not just CallWare, I have added some more initials after my name - CMTP, for Certified Multimedia Telephony Professional. Yes, I know, that acronym sounds overinflated and a bit too self-important. I liked it better when it was just called CNTE, Certified Network Telephony Engineer, since that sounds more to the point.
Naturally, I think it’s a great product, or I wouldn’t have wasted my time and money on the training. Since this is obviously no longer an unbiased opinion, I won’t go into the details of this product in this article. I just wanted to mention the result of the research I’ve previously written about and didn’t want to censor information about a quality product just because I’ve decided to sell it.
Windows 97
Windows 97 (or, as I like to call it, New Windows Version 1.00, as opposed to Windows 95, which I call New Windows Version 0.95), it has been delayed again. It is now being called Windows 9x, and, since its new due date is the first quarter of 1998, will presumably be renamed Windows 98. (If it ends up named Windows 99, I suspect the software development industry as a whole will rise up and lynch Bill Gates.)
The previous due date for this product was August 1997. This is a major disappointment for me since I expect it to have the same relationship to Windows 95 as NetWare 4.1 did to NetWare 4.0 - that is the real, working version of what amounts to a beta release to the public.
I would refer Mr. Gates (and anyone else who wonders why software projects are seldom delivered on time or bug-free) to the research done by IBM in the 1960s covering the original automation of the New York Times, along with associated research from that period regarding NASA’s Agena project.
In its research IBM found that at some point, throwing more people (and presumably money) at a software development project becomes counterproductive, due to the intricacies of multiple person-to-person communications procedures. As you add more people past that point, the project actually slows down and will eventually grind to a halt!
The real limiting factors turn out to be organization and planning. Software development teams that suffer from an autocratic leadership and bureaucratic interference cease to function effectively.
Translated into real terms, it means that if management changes plans every few days, don’t expect to make your expected release dates or to deliver reliable software. Those who best balance the competitive demands of the computer industry with the practical necessities of the development process create reliable products that are delivered on time. Those who base their sales mostly on overwhelming marketing spending are likely to deliver late, buggy products.

�1997, Wayne M. Krakau