Top White Papers
Rant Mode Equals One: The Fragmentation FearNov 23, 1999, 20:44 (18 Talkback[s])
(Other stories by Paul Ferris)
By Paul Ferris
[ The opinions expressed by authors on Linux Today are their own. They speak only for themselves and not for Linux Today or internet.com. ]
I keep hearing the fragmentation scare, over and over. I keep hoping that the clue meter will read above zero for a few minutes and the whole issue will quietly die a deserved death. It's one of those basic fears concerning Linux that has come about because a lot of people associate fragmentation with the death of Unix.
Here, in order to simplify things, I'm grouping things like Solaris (from Sun) with HP-UX (from Hewlett Packard), AIX (from IBM) and other competing Unices with names and owners too numerous to list into one group product called "Unix".
First, I'd like to address the so-called death of Unix. Since the combined Unices power a good portion of the web and the infrastructure of many Fortune 500 companies, it's hardly "dead". Sure, sales are a bit flatter than those of proprietary commodity desktop operating systems such as Windows NT, but who's to say that it's a dead platform just because the particular market segment it occupies is different?
Semi-tractor-trailer rigs should be classified as "dead" for the same reason. Their market growth is flatter than those for sport utility vehicles (SUVs). Maybe we should stop transporting all of the bulk items that are conveyed via semi-tractor-trailer rigs and move them to hoards of people using SUVs. Heck, the training for operators of those big trucks is extremely expensive anyway. Imagine the cost savings when we replace those individuals with untrained legions of droids driving SUVs.
In case you believe this fear, uncertainty and doubt about the semi trucks being dead, try stepping out in front of the next one you see driving along the road. Some Unix fanatics and I will be glad to watch and take a vote over just exactly what "dead" means.
The reason people want to make this comparison with Unix centers around the fact that Linux is similar to Unix. Since Linux operates like Unix from a design and interface point of view, people want to compare it in all other aspects as well. But Linux is not Unix in a very fundamental sense.
The traditional Unices each had different companies that were responsible for their creation and growth. The source code used to create these products was proprietary and the property of their respective owners. This created many different variants and extreme competition among rivals. The differences resulted in life being difficult for software developers who had to make different versions of their product for each platform. This was tricky and Microsoft promised to simplify this problem by creating only one code base for it's Windows NT product. NT was to supplant these traditional Unices and run on their respective hardware platforms.
Ignoring the fact that competing interfaces and standards are a good thing for progress, this fragmentation was a bad thing from the viewpoint of the software developer and therefore it was a bad thing for Unix. Now people are addressing this same crowd, and predicting a similar fate for Unix's new evolved cousin, Linux. These people are overlooking something rather important: The source code for Linux is open and in a place where everyone can see it.
Although it could happen that the code base actually "fork" or diverge into different products, even if this did happen, everyone would be able to see exactly what had changed about the application interfaces. Everyone would be free to mix those same pieces of code right back into the opposing code base. It is ignorance of the licensing scheme of the Linux kernel that is the real problem.
What people want is the reassurance that there will be a single point of control for Linux. That's the way it is for legacy systems like Windows NT. It's the idea that a lot of people are used to with products that originate from a corporate realm.
My advice to these people is simple -- stop trying to impose restrictive control methods that squelch true competition and evolution upon something as dynamic as the free software development model used by Linux.
Windows NT's strength as a unified proprietary product from a corporate entity is also it's Achilles heel. Look how many different hardware platforms are available for Unix and Linux and compare that to exactly one for Windows NT. Look at the slow development time for Windows NT, er Cairo, er Windows 2000. Look at the cost factors which continue to rise.
Ask if the promise of lower cost and a single API have actually been delivered in the face of Windows 2000's new pricing scheme and the fragmentation of the APIs provided by WinCE, Windows 2000 personal, Windows 2000 professional and 64 bit Windows.
Microsoft has not proven that they had a good idea "unifying" their code base in this proprietary manner. What they have proven is that they are just as bad or worse than the previous model. At least previously companies had choice in the hardware area.
Linux is doing just fine today and the fragmentation that is the big fear still has yet to rear it's ugly head. As Eric Raymond recently pointed out, traditional Unices are actually starting to unify around Linux.
I'm sick of fears being based upon no evidence what-so-ever. It's pathetic. When will this kind of mindless parroting of baseless ideas stop? I can understand the fear, it's based upon Unix -- but I cannot understand why people haven't noticed the major difference between Linux and Unix -- the development model. The difference is huge and the fears unfounded.
There is a paradox here and I'm sure it confuses the issue.
Linux is already fragmented -- there is no central point of control, other than the final say of Linus Torvalds. But at the same time it's not fragmented -- programs work because the API is open and no one can produce code bases with hidden source code. It's not allowed by the licensing scheme.
Take your pick, but please don't spread panic about it happening. Either it has already happened and it's all just fine that way or it's not ever going to happen so it's no big deal. Take your pick. Stop riding the fence and shouting about things that can't happen.
I know I'm wasting my breath here, but people continue to parrot this fear and it's getting old. I've decided it's time to try another attack -- give people some real fragmentation fears.
Scary "fragmentation" predictions by Paul Ferris for the coming millennium:
The automotive design market will fragment into various segments and brands. Several different brand names will emerge and interchangeability of parts will become a pipe dream. Each different brand will require it's own mechanical specialist to work on it. Somehow, despite this enormous setback, people will still be able to find and use quality cars and easily find places to service them.
Christianity will fragment into two groups: Protestants and Catholics. Although some of them might get into pissing matches over which one is better at times, the religion will continue to be practiced by a huge portion of the world.
The human race will fragment into two genders, male and female. The two genders will tend to do a lot of things differently and poke fun at each other often. A lot of interesting interoperability will be achieved in spite of it all.
The United States government will fragment into warring factions. Some people will call themselves "Democrats", others "Republicans", still others "Independents" and "Libertarians". Oddly enough, it will still function.
Maybe Bill Gates and Microsoft can clean all this up with some homogeneous proprietary software when it happens. All those different people could learn to salute the Microsoft flag and just listen to good ole' Uncle Bill.
The day will fragment into two vastly different experiences. Part of it will be dark, and the other part light. The sun will rise in the East, and set in the West.
I'm willing to bet that a lot of my predictions here will come true.
And, I'm hoping now that the fragmentation issue is very, very dead.
Rant mode equals zero.
Have a nice day.
Contribute your fragmentation worry or prediction below, I'm quite sure that I was too busy ranting to get them all....
Paul Ferris is a husband, father and the site developer for Linux Today. His computer expertise spans over 13 years of Unix, VMS, and various PC operating systems. An avid Free Software fan, he has helped support Linux since 1995. His pastimes include Linux, science fiction, metaphysics, spirituality and classic automobiles.
0 Talkback[s] (click to add your comment)