What you don't know can hurt you
May 9, 1998
by Zac Belado
Early indications point to Windows NT5 being a potential source of very serious problems for IT managers and system administrators that will be required to install it when it rolls out sometime in the next millenium.
The second beta of NT5 has yet to be released and already consultants like the Gartner Group (http://www.gartner.com/) are suggesting that their clients wait until at least 2001 before moving their systems to NT5.
Let's look at that statement again. A major PC consulting firm is recommending, before even seeing the final product, that their clients not initially use it. They are, one would have to assume, expecting that the initial release of the operating system will contain enough serious bugs to make using it, to be polite, a less than optimal experience.
They're not bugs. They're sub-optimal features
Part of the problem has to be the very nature of Microsoft's business plan. Two very important factors cause Microsoft's software to be a constantly moving technology target. The company makes a very sizeable amount of its money on upgrades. As well, the company constantly adds new features to its operating systems (and announces them far in advance of their actual arrival or even production) in order to keep clients in the Microsoft "orbit".
These two "forces" are inter-related and tend to feed off each other. Features are added to products in order to drive (and some would say keep) sales and the need for new upgrades and sales are what push early release of products that are not yet completed. The recent release of Windows 98 illustrates this feedback loop perfectly. A product gets shipped far too soon in order to ensure that sales are generated.
Windows NT5 is less a growth from NT4's codebase and more a reaction to UNIX. It is reported that the codebase of NT 5 will contain somewhere in the order of 35 million lines of code. Compare this to NT 4 which, with all the latest service packs, has somewhere around 15 million lines of code.
Ponder the numbers for a second. Microsoft programmers are adding the equivalent of an entire new OS on top of NT 4.
And they still are fixing the bugs in NT 4.
So what the heck are they adding?
I recently spent a bit of time looking around the Microsoft web site looking for a detailed feature list for NT 5. I assumed for quite a while that one didn't exist until I came to the realisation that the "overviews" I was reading were the list I was seeking. While I still haven't seen a detailed list of the proposed new features for NT 5 (http://www.microsoft.com/ntserver/basics/future/windowsnt5/default.asp is a good place to start) nothing I have seen so far seems to add up to 20 million lines of code.
A very good answer to the question of what they are adding is available at the NC Magazine web site. In the last "Last 10 Minutes" column (http://www.ncworldmag.com/ncworld/ncw-06-1998/ncw-06-lastten.html), the author Nicholas Petreley provides that answer by, in part, positing that "Microsoft will deliver on its promised features set but Windows NT 5.0 will rank among the greatest programming disasters in American history".
The short answer to our question is that they are patching the NT 4 codebase in order to provide the functionality they promised to keep their clients from migrating to UNIX. Petreley's article has an excellent synopsis of the numerous problems that Microsoft faces in trying to out-UNIX UNIX. Creating, in effect, a way to maintain a desktop oriented operating system and try to leverage it into a distributed network model.
Or put a square peg into a round hole. Which, as you may know, can be done put requires you to add serious modifications to either the hole or the peg.
Hence an extra 20 million lines of code.
Where do you want to GPF today?
Which might be why the nice folks at the Gartner Group are so leery of suggesting to their clients that they use the initial release of Windows NT 5? And frankly Microsoft's recent track record doesn't really do anything to give anyone any hope that it might be otherwise.
Some recent examples:
- Windows 98 shipped with a bug that allowed anyone (with knowledge of your IP address) to waltz through your machines local drives if you had file sharing turned on (Quick Quiz for Windows users. Is file sharing enabled on your machine? Do you even know where to look?).
- NT 4 had a bug in its file system that caused it to deliver scripts (like ASP and Cold Fusion pages) directly to a browser without first parsing the files. Leaving all sorts of delightful information like database paths and passwords open to casual viewing. I spent the better part of two hours viewing the ASP code from Microsoft's own site.
- The Back Orifice system, created a group of "underground" programmers called Cult of the Dead Cow, which effectively opens entire Win95 and Win98 computers to attack is described as being made possible as a result of systemic security failures by Microsoft.
Can 50,000 Elvis fans be wrong?
A friend of mine was born and raised in Yugoslavia. One day during lunch he related some of the deficiencies of the command economy in Yugoslavia. How companies were run in a haphazard fashion, related industries didn't co-operate, sales staff were rude and abusive and how the entire system actually effectively rewarded inefficiency.
Now let's compare this to Microsoft. A company that is more interested in selling you a Microsoft product regardless of its quality. A company that has effectively destroyed any competition it might have had. (As an immediate illustration of this point, try to find a copy of PC Computing from six or seven years ago and notice the number of products that were available then). Microsoft has become the computer industry's "central committee" in an effective command economy that starts and ends with Microsoft.
The last major competitor that Microsoft faces is the numerous UNIX clones that run most of the mission critical systems in the world. And that is what NT5 is targeting.
And that is why it is so dangerous.
Copyright 1997-2024, Director Online. Article content copyright by respective authors.