On October 31st 1998 (Halloween) Eric Raymond posted a leaked memo from Microsoft. You can see it (and follow ups) at <URL:http://www.opensource.org/halloween.html>.
This prompted Jordan Hubbard to write the following response. This text is unchanged, except for the HTML formatting.
[ caution - this is a bit long. Lots of points here I've been wanting to cover for awhile and now seems as good a time as any.. ]
OK, so we've all seen this latest bit of Linux leaping about and shouting from the rooftops and some of us have even gone "agh!" and run around a bit ourselves, but now that we've all hopefully calmed down again I'd like to say a few words about this and the state of FreeBSD advocacy in general.
First off, just to cover the Halloween memo in brief, yes it appears to be genuinely from Microsoft and yes, it appears to be genuinely full of statements culled from various Linux evangelists who feel no pangs at making blatantly false pronouncements like "Linux is the only OS experiencing growth" or "Linux is the only contender for the x86 platform." These types of statements are pure hooey, of course, and FreeBSD is currently doing better than it has at any previous point in its history. Our releases are starting to finally hit their stride, it seems (and try to remember back to the days when it was more like: "My god! We did it! A release!"), and our rate of innovation and self- improvement hasn't been higher since the 2.0 days - it's very encouraging to see that we can spur ourselves to such heights of productivity *without* legal injunctions staring us in the face! :-)
Second, we have to keep sight of the fact that none of this is particularly new or even interesting. We know that Linux is the current poster child of the press and we also know about the press's irritating predilection for focusing on one and only one champion rather than looking more in depth at the situation. We can yell and scream all we like, but we're not going to change the fact that for many journalists investigating "Open Source", Linux is the first and possibly only thing they're going to look at. It simply has the right sized hype-bubble surrounding it where we do not. We also have to accept the fact that ISVs are going to target their products at the much more obvious Linux market and try to strike deals with it, going "FreeBSD? What's that?" when asked about a native port. The same goes for investment, selling shares in Red Hat, Inc., etc. Money always goes after the visible markets first.
What you have to ask yourselves, looking at the dynamics of this situation as dispassionately as possible, is whether all of this is necessarily as bad a thing as some of the gloom-n-doomers would have us believe. Looking at only the superficial indicators, it's easy to say that "Linux is winning and we're losing", pointing to the stacks of Linux books and magazines in the bookstores, the Clinton transcripts where he mentions Linux, the Goodyear blimp circling overhead with Linus's smiling face shining from it, etc etc. It's especially easy to say that when you hold Linux and FreeBSD in your mind as equivalent products, started at the same time and with the same overall development mentality.
The fact of the matter is that Linux and FreeBSD are NOT equivalent products with identical user and developer communities surrounding them, however. We've *always* been lower key about things, preferring to quietly focus on the business of steadily turning out quality products to only moderate fanfare. It's no use screaming for teams of FreeBSD fan dancers to come out and start singing the praises of FreeBSD in full 4-part hyperbole with some grinning, cigar-chomping promoter standing in the background - that's just not us. The nay-sayers will also say that "this not being us" will surely be our downfall since you gotta sing and dance now if you want to be noticed, but I'm really not so sure about that. To my way of thinking, we have our style and we have our niche and they're both respectable in their own way. Not everyone buys toilet paper because a team of singing rabbits (to paraphrase the great Rod Serling) suggested it on television, and some people DO react positively to the somewhat less superficial attributes of quality, consistency and a focus on the technology rather than on standing in front of the cameras and saying things like "open source validates the concept of a basic human sociological tropism towards cooperation and the free and open exchange of .." to some vapid blond on Technology Week.
That kind of approach might also get all the sound bites this week, but remember the old "15 minutes of fame" effect and the fact that the press is going to get bored with Linux eventually and go off in search of other things they don't understand to dissect. When that inevitably happens, it's going to be back to quality and those groups who remained true to their basic operating principles and didn't get sucked in and destroyed by excessive growth or hype. The opportunities for wandering off and getting lost in the woods in pursuit of some bright and shiny object have never been higher than they are now, and somebody's bound to panic and go off and do something stupid in an effort to differentiate themselves. I don't think we have any need to panic at all and should simply keep doing what we're doing and try to do it as best we can.
I'm not saying that there's no room for improvement, and some alliances *are* being made with various artist/marketing types whom we think can help us get the attention we deserve, but it's not the same as saying that we're going to drop everything and go play Linux's game now. That would be the wrong move and I can only point to the history of BSD itself when searching for good examples of technologies which have remained viable long after "losing" a war to a competitor. BSD "lost" to SYSV over a decade ago, but did that kill it? Quite apparently not and it appears to be doing better today than it ever did even back in its heyday, when it ran on a large collection of VAXes but hardly any of the commodity (68K) hardware at all (you had to buy an obscure 32016 based machine if you wanted to run BSD at home :-). The situation today is vastly improved by comparison and most people don't even stop to think about that.
In any case, I didn't mean this posting as a fluffy "we're fine!" sorta thing, though I do think that people sometimes lose sight of our own growth rate and notable successes when furrowing their brows over the latest Linux PR victory, I do have a summary of points I think we can and should improve:
Keep pushing the magazine articles out. These seem to be easier for people than books and I've largely given up on trying to incite a FreeBSD book to happen - I guess that will just occur in its own good time. Walnut Creek CDROM is still paying a bounty for magazine articles (matching funds for your fee) and has enabled more than one person to buy a new machine for the price of one weekend's writing for a good cause. Pick a target publication and go for it, folks! I've done about 3 of these so far (maybe more, I forget :) and can say that it's not that hard. You generate a simple article outline and you submit it to the editor along with your proposal for what you're trying to accomplish with the article (just a paragraph or two of text, not a thesis). If they're interested, they'll send you back details on how long they want the article to be (generally 500-1000 words) and how much they're willing to pay. When they pay, send us a photocopy/FAX of your royalty check and we'll pay too. It's that simple, and it good for FreeBSD to appear in print like this since it reaches outside the somewhat closed audience of the mailing lists.
Look at Linux as a door opener, not a threat. I mean this, folks, even you rabid Linux haters out there. Consider very carefully the fact that if customer A needs a PC to do server job B, customer A is going to do one of four things:
- Buy NT
- Buy a commercial Unix
- Buy Linux
- Buy *BSD
Those really are about the only 4 options for building a department fileserver or gateway box with cheap, commodity hardware (we'll assume the people who don't want cheap buy Cisco gear, Suns and NetApp filers anyway) and let's look at them in turn:
If they buy NT, you can pretty much write them off. By the time they realize what they've gotten themselves into, the investment (or embarrassment) is generally too great to back out of anyway and it's actually very few IS shops that seem to claw their way back from NT and install a free OS instead. Sure, you hear widely trumpeted stories whenever some large ISP does make it back from NT, but its very rareness is what makes it something to trumpet about. NT is Darth Vader here and we must fear his control of the dark side (marketing) and the fact that "everybody knows NT" when the issue of personnel comes up with most pointy-haired managers.
Is a much better option since at least the customer has accepted Unix as their savior and can potentially be won over at some point by OSS, but the fact that they chose a commercial Unix probably also means that they have deep-seated needs for tech support or inter-operability with other parts of the IS shop and you'll probably have to work on them for awhile to win them over.
Here now we've at least accomplished two things: We've got the customer admitting that they want Unix and that they want a free Unix. Furthermore, they've chosen a solution which we think we can beat in all the taste tests if we can just get the CD in front of their faces. All in all, this has got to be the easiest conversion of the three and a definite win if their only other options were A or B.
Yay! Of course we like this one, but if it's not FreeBSD then we still have a bit of a conversion job to do and it might even require something like a SPARC port to be able to offer the same cross-platform inter-operability that the user has chosen the other *BSD for. It's something to think about, and certainly no better than the Linux scenario in some ways (again, if you're just thinking about this from the pure, mercenary "how do we get more FreeBSD users" perspective).
Hold your advocacy to a higher standard, and by this I mean that if we're to weather this whole PR blitz period with our reputation for being "the calm and level-headed ones" intact, we can't stoop to the level of some Linux advocates when trying to make short-term gains against their PR blitzes. Sometimes you just have to be Gandi.
When the press have gone away, believe me, people will remember which groups stuck to their guns and didn't compromise their identities or ideals and which went sort of nuts and participated in a few raping and pillaging sessions. I'd far rather be the group still standing there when the smoke clears going "Yup, we're still here and still doing good software without the fanfare or fancy costumes. Have a look!"
To put it another way: If FreeBSD were a respected musical entertainer, I would want her to be the one who stuck to doing the kind of music she liked and always did it well rather than horrifying us during the disco years by suddenly putting on spandex pants and lip-syncing to formulaic, song-factory material or shrieking out heavy-metal lyrics in heavy makeup with Axel Rose 10 years later. :-) Sometimes the price of "success" is too high.