Steve Hargadon just came over to me. “Go look in the session room.”
“How come?” I says.
“Just go look.”
Turns out that the Moodle session is going on in the Open Source session room — and it looks like a Who concert. People sitting and kneeling at every table, and sitting and standing at the back of the room. Not a square foot of empty floor space in the entire room.
There comes a time in any given IT market when the “open source” switch goes off in people’s minds. Education is clearly at that point. And no one is so zealous as the newly converted. It’s a very exciting time.
I wish I’d brought my camera. Oh well.
I’ve always been a big fan of Coming to America. To me, it represents in many ways the artistic pinnacle of Western achievement. One reason: there’s a parable in Coming to America for just about everything.
Today’s parable comes from the end of the movie. Maybe some of you remember the scene:
“Wait a minute, wait a minute, wait stop right there! Listen: Stop right there, man. A man goes into a restaurant. You listenin’? A man goes into a restaurant, and he sits down, he’s having a bowl of soup and he says to the waiter, waiter come taste the soup. Waiter says: Is something wrong with the soup? He says: Taste the soup. He says: Is there something wrong with the soup? Is the soup too hot? He says: Will you taste the soup? What’s wrong, is the soup too cold? Will you just taste the soup?! All right, I’ll taste the soup – where’s the spoon?? Aha. Aha!”
It’s funny when someone comes up to you on the showfloor and *makes* you taste the soup. For me, today, it was this question: “Can you show me how to remove packages without using the command line?” So I brought up Pirut on Fedora 7, and because I didn’t have network, it wouldn’t even load. “No networking? No package list. Sorry.” And when I tried it on RHEL5, it took me about 5 minutes of staring at the GUI to realize that unchecking a box removes a package.
“Yeah, yeah,” my inquisitor said to me. “You don’t even notice, I know, because you would go the command-line to remove a package. But my typical user is a science teacher who became the technology coordinator for her high school. And I’m sorry to tell you this, but she will never bring up a command line, ever.” He then challenged me to manage my system for a month without using the CLI at all.
I think I’ll take him up on that. I think a lot of people in Fedora-land should. And we should be tasting not only our own soup, but the soup of others, too. Here’s the thing: everyone’s soup is basically the same, and we keep saying “what? Our soup and their soup is equally tasty. Is our soup too hot? Is our soup too cold?” But smart users are pointing out to me, more and more, that sometimes they can’t really taste our soup because we’re trying to feed it to them with a fancy salad fork.
Ah, what you know from funny, you bastards.
But first, let me tell you why I care. Everytime you see “the outside” or “the margin”, think of the Fedora community. And everytime you see “the inside”, think of Red Hat engineering.
One reason so many good ideas come from the margin is simply that there’s so much of it. There have to be more outsiders than insiders, if insider means anything. If the number of outsiders is huge it will always seem as if a lot of ideas come from them, even if few do per capita.
Outsiders should realize the advantage they have here. Being able to take risks is hugely valuable. Everyone values safety too much, both the obscure and the eminent. No one wants to look like a fool. But it’s very useful to be able to. If most of your ideas aren’t stupid, you’re probably being too conservative. You’re not bracketing the problem.
The very skill of insiders can be a weakness. Once someone is good at something, they tend to spend all their time doing that. This kind of focus is very valuable, actually. Much of the skill of experts is the ability to ignore false trails. But focus has drawbacks: you don’t learn from other fields, and when a new approach arrives, you may be the last to notice.
So if you’re an outsider you should actively seek out contrarian projects. Instead of working on things the eminent have made prestigious, work on things that could steal that prestige.
Outsiders are free of all this. They can work on small things, and there’s something very pleasing about small things. Small things can be perfect; big ones always have something wrong with them. But there’s a magic in small things that goes beyond such rational explanations. All kids know it. Small things have more personality. Plus making them is more fun. You can do what you want; you don’t have to satisfy committees. And perhaps most important, small things can be done fast. The prospect of seeing the finished project hangs in the air like the smell of dinner cooking. If you work fast, maybe you could have it done tonight. Working on small things is also a good way to learn. The most important kinds of learning happen one project at a time. (“Next time, I won’t…”) The faster you cycle through projects, the faster you’ll evolve.
It strikes me that there are lots and lots of small things that could be built with Fedora, or on top of Fedora. And the more I think about it, the more I’m glad that we build Fedora with this stuff in mind. And the more fervently I hope that folks in Fedora land decide to do the “risky” things that Red Hat refuses to do.
…to the horde of Twitterers.
Meanwhile, I spend my Saturday afternoon setting up demos for NECC 2007. OLPC, Fedora 7/Revisor/Live CD, a bit of RHEL5 desktop, all kinds of good stuff to talk about. Shawn Briscoe, Red Hat education czar, has asked me to wear my red fedora — which I don’t generally do, but at a big trade show it’s an incredibly effective way of getting people to talk to you.
I’m excited. There’s something happening in education right now. One-to-one computing is an important movement in the educational computing space, and we’ve really got a lot to offer there on a lot of levels. So I’m sure I’ll meet a lot of interesting people and I’ll learn a lot of interesting stuff. It’s really not hard to install Linux in an entire school, and get as much benefit or more as you can from an equivalent Windows purchase — for a lot less money. The key question: how do we build a strong network of professionals to support these school efforts the way they need to be supported? That’s what Shawn and I will be trying to figure out this week.
Also, I guarantee that *dozens* of teachers will be asking how they can get OLPCs in their American classrooms. My standard answer is “once we know it’s actually gonna work.” But I think it’s time to start developing a more coherent answer. If anyone has a good one, I’d love to hear it. :)
…it’s possible that I was too harsh in my assessment of the ZFS licensing situation.
It is possible that Sun fully intends to put ZFS, in its entirety, under both CDDL and GPLv3 — and the reason Jonathan isn’t saying so is because his lawyers threw a hissy fit when he proclaimed Sun’s intention to similarly dual-license OpenSolaris before the lawyers were “ready”.
I should do well to remember: never attribute to malice that which is adequately explained by an army of nervous lawyers.
So we’ll see. Maybe ZFS will be dual-licensed under GPLv3 after all. And if so, then the ball really is in Linus’s court.
Jonathan Schwartz says that Sun is “interested in seeing ZFS everywhere, including Linux, with full patent indemnity.”
If this is so, Jonathan, then tell us: will you be releasing ZFS under GPLv3, in its entirety? Because if you say yes, that basically ends all debate immediately, doesn’t it?
Hey, you can even keep it under CDDL as well, so that there’s no problem for those who want to link ZFS to non-GPL binaries. Since you are the copyright holders for the whole shooting match, it should be a simple matter to dual license ZFS under both CDDL and GPLv3. Right?
So let’s keep our eye on the ball, folks.
If Sun is serious about making ZFS available to all Linux distributions, then they will, at some point, dual-license ZFS — in its entirety — under GPLv3 and CDDL. If that happens, golf claps all around.
Let’s assume, for the sake of argument, that this does not happen. (And since Sun developers have basically said as much, I think it’s a reasonable assumption.)
So if it doesn’t happen, then one must ask the question: why not? Why might Sun choose not to dual-license ZFS under GPLv3 and CDDL?
This question is especially relevant when one considers that Sun has chosen to release Java, OpenOffice, and OpenSolaris under GPL or GPL+CDDL licenses.
We must thus infer that Sun sees some fundamental differences between the licensing implications around OpenSolaris/OpenOffice/Java and ZFS, right? And we must infer that such differences correspond to the licensing differences between GPLv3 and CDDL, right?
So what inferences might a reasonable person make?
Well… um… it’s patents, right? I mean, what else is there?
We all know that Sun has been very active in protecting ZFS through patents — 56 patents, according to Jeff Bonwick of Sun.
The most notable difference between the CDDL and the GPLv3 is that the GPLv3 has very explicit anti-patent language. The gist of this language: “no discriminatory patent practices around GPLv3 code.” The CDDL also has patent language, but it’s not nearly so anti-discriminatory; read section 2 of the CDDL for yourself, and look very closely at those cases where patent rights are explicitly not granted to licensees.
Thus, my inference: I personally suspect that Sun wants to reserve the right to assert, and thus cross-license, its ZFS patents selectively. CDDL allows them to do this. GPLv3 will not. Therefore, they will not dual-license ZFS, in its entirety, under CDDL+GPLv3.
Which means that Linux will not — in fact, cannot — incorporate ZFS. And the folks at Sun will continue to say, “look at how unreasonable the GPL is,” even though they happily put the GPL to work in all of their other projects. And then Sun folks point to the great work that FreeBSD is doing with ZFS… whereupon Theo punches Jonathan right in the mouth for being “duplicitous”. How’s that for gratitude? Heh heh heh. Those FreeBSD guys are feisty.
I could, of course, be completely wrong about all of this. I would be legitimately happy to hear someone from Sun tell me I’m wrong, and then prove me wrong by dual-licensing ZFS under GPLv3 and CDDL. Then everybody’s happy, and ZFS makes its way into Fedora 9, and the whole open source community grows and prospers, hurrah!
See, Jonathan, despite all the good stuff you’re doing at Sun, this is the kind of stuff that continues to make everyone suspicious. You say “remember, we can’t put ZFS under GPLv3 because GPLv3 doesn’t even exist yet,” even though you’ve already proclaimed that it’s precisely what you’re going to do with OpenSolaris. Why? Because it suits you to say “OpenSolaris will be GPLv3,” and yet somehow it does not suit you to say “ZFS will be GPLv3″, for some reason that you can’t quite explain. “Indulge me,” you say, and then you seem surprised and hurt when people like Linus (and Theo, heh) don’t.
ZFS under GPLv3+CDDL is a perfect example of something that should just happen. It works for everybody, it makes all the code as free as free can be, and because you own all of the code, you can just snap your fingers and make it so! But you don’t. And then reasonable people make reasonable inferences that you can’t quite reasonably refute.
Meanwhile, Linux keeps on getting better and more free. And it’s the GPL that has always made it so.
(p.s. IANAL, as though anyone needed a reminder of this.)
It’s a fact of life: awards, no matter what form they take and no matter how well intentioned, leave disappointed people in their wake.
Michael Schwendt has expressed his disappointment at not having received this award, and now I find myself wondering how many other invaluable Fedora contributors say nothing, but silently feel hurt that their contributions were not judged “worthy” for this particular recognition.
Michael, to you in particular — and, by extension, to others who may have felt similarly but didn’t want to say so — I sincerely apologize if this award made you feel unworthy in any way. I can assure you, making you feel like crap was the absolute *last* thing on our minds. Why didn’t you receive this award, when others did? To be completely honest, I don’t remember. We came up with the names, we sat on them for a while, and when we released Fedora 7, we announced the winners. It wasn’t much more complicated than that. That’s not a very transparent process, and not very satisfying to “winners” or “losers”, I’m sure.
Here’s a humbling stat. Google “Greg DeKoenigsberg” and Fedora: 25,400 hits. Google “Max Spevack” and Fedora: 52,800 hits. Google “Michael Schwendt” and Fedora: 96,800 hits. More than me and Max combined.
Sometimes an idea is better in theory than it is in practice.
So let me ask this, friends and Fedora contributors: should we reward the “best” of the Fedora contributors at all? And if we do, how should we do it?