Features

net.wars: White hats, black hats, who's got the grey hat?

So the PR guy asks me, he says, "Where do you stand on the whole Microsoft versus Linux thing?"

I say, "It's all on my Web site."

He doesn't believe me, but I say it again and away he goes to investigate. He does not call back, so I can only assume he has found the lengthy archive of net.wars columns. Actually, none of them talk particularly about "Microsoft versus Linux", but the alert, slightly paranoid, PR reader might detect hints.

Had he persisted, I suppose I would have said, the way journalists do, "I think Microsoft is the Evil Empire and Linux is the Messiah that has been sent to save us."

A really smart PR at that point would laugh and say, "No, really, what do you really think?" (A dumb one would suddenly discover that there was actually no room left and he was awfully sorry they couldn't accommodate me this time.)

Stupid, meaningless question

And at that point I'd tell him: I think it's a stupid, meaningless question. It reduces a complex set of issues into an absurd cowboys and Indians game. There is no "Microsoft versus Linux" question. There is a genuine need to understand what parts of the computer industry are best served by proprietary software and what are best served by open-source. The first may or may not be made by Microsoft; the second may or may not be GNU/Linux.

The best essay I remember reading about this came from Clay Shirky, whose essay The Interest Horizons and the Limits of Software Love talks about this very issue, partly in response to Eric S. Raymond's famous essay, The Cathedral and the Bazaar. Certain kinds of software, he argues, like printer drivers and accounting software, are "unloved". They are not the kinds of things a load of volunteers find rewarding to write. They're not fun, and what is important about them is reliability, not creating something cool your friends will admire and that does something you've wished software would do for a long time.

The road: open source technology

So a better question for the PR person to have asked would be, "What kinds of software are best made under a proprietary regime versus using open-source methods?" And, "what model best encourages useful innovation?" Useful is the key word: when I pointed out that the open-source Sendmail and have run vital bits of the Net for 20 years, the PR guy said, "Ah, but what innovation has there been?" Innovation? We just want those bits to WORK INVISIBLY, dammit. When you drive on a 50-year-old road, which is a nice bit of open-source technology, do you say, "Where's the innovation?"

A more interesting take on where software might be going came from last weekend's Notcon, where Danny O'Brien delivered a talk on Life Hacks. As part of his research, he surveyed a load of highly productive programmers to find out how they become that way. One of his questions asked them what secret software they used. These are the bits and pieces they smash together to fix something that drives them crazy. They'll clean it up someday so other people can have it. You know, like that shirt you only wear around the house because it's missing a button.

Surveying this stuff ought to be gold for a commercial software house. O'Brien is finding patterns: everyone has a utility to kill whatever program regularly turns into a memory hog; everyone has a scraper to pull down the data from their online financial services quickly; everyone has utilities for email wrangling. But they're all Linux hacks or run in Perl, things that the average Windows user never touches. Geeks, as Ellen Ullman has observed, are the canaries in the mine. What they are frustrated by today, we will all be frustrated by in a few years. Innovate this way.

Which leads back to the main point. All the Microsoft speakers directly compared Total Cost of Ownership between Windows and Linux in various contexts. They didn't provide a lot in the way of facts and figures - as the best propagandists don't - though they did refer us to their Web site on the subject. But the big point about open-source software is, as Richard Stallman keeps saying, "free" as in "free speech", not "free" as in "free beer". The right to inspect, modify, tinker, improve, fix, and add to the software you use may outweigh simple cost comparisons in some - not all - contexts. Shared source, which Microsoft says is an idea it did copy from the open-source movement, only gives carefully chosen organisations the right to do this. It is not the same.

Indemnifying code without indemnifying the user

Ah, they

said, but you see we couldn't then indemnify the code. Oh? I am, I said, unaware of any situation in which Microsoft makes good damages to a company if its software fails at the wrong time. That's not what they mean by "indemnify". They mean: they couldn't guarantee that all the code in Windows was valid (insecure, but valid). Besides, intellectual property is all we have in the way of corporate assets, other than maybe our people.

Maybe our people? Join the open-source movement!