Flutterby™! : On Free Software 2002-01-21 19:41:39+00

On Free Software

2002-01-21 19:41:39+00 by Dan Lyke 3 comments

Charles called me up for some quotes on Linux for an article he's writing. The sound-bite I should've given him:

What high-level decision makers should realize is that Linux is already being clandestinely deployed in their server rooms, not for cost reasons, but for functionality reasons. Open Source may actually cost more than proprietary software if you're customizing it and adding patches back to the code and using it in the spirit that it's written. But the point is that you can't save your way to profitability; at some point you have to be able to do things better and quicker and more reliably than your competitors; cheaper just makes you the next KMart.

I agree with your comments about the cost of Linux. What you save in cash outlay you more than pay for in fiddling around with it. Time is money. If this gives you more functionality, so much the better. However, even if the cost/benefit analysis must be rethought, it cannnot be ignored. Too many of the Linux true believers in discussions like ZDNET seem to feel that the entire world is waiting with baited breath for the possibility of having a file manager from one designer, an interface from another, and a graphics package that has to be rebuilt everytime you need a new color. Our here in the real world, nobody has the time, much less the expertise to screw around with the OS all the time. I think this attitude is one reason the desktop end of the business is still largely closed to Linux. Its great niche will remain in the server and other high performance markets, where the systems have the people and the time to support them.

As to Kmart, the reason Walmart can consistantly beat them in price is that they spent years developing a just-in-time inventory system that is the envy of the industry. Vendors wishing to sell their products in Walmarts must integrate their delivery systems with Walmart's. Your right, selling price is only half the question.

I actually don't think "pay for in fiddling around with it" is a real cost. My experience is that frustrated time spent adminning machines is roughly Mac, Windows and Un*x variants, worst to best. The issue there is that anyone can sit down in front of the Mac and do simple stuff, it's the complex stuff that's more

difficult than on a Un*x box, but getting over the initial hump on a Un*x box is harder. Thus a professional admin is going to be far more productive on a Un*x box (and most studies of cost/benefit reflect this).

The real issue is that unless you are rolling a lot of your own code, you are not the customer for open source software. "Free" software is not immune to market pressures, it's just that the currency is different. We talk about "user oriented", and the point is that open source software is entirely

user oriented, it's just that it's oriented towards the same types of users who want to write it.

So if you're not willing to budget for putting code back into that pooled base, and budget about what you'd spend on commercial software, you're not going to reap the increased reliability, speed, and similar benefits. This is also the reason that the recent Gates memo on security can only just be PR: Microsoft's customer base has always

been willing to trade ease of use for security; real security starts with education of users, and if you switch that responsibility you'll alienate all the people who prefer the Microsoft mode of interacting with their computers.