OSCON interview snippets
OSCON 2009 is now drawing to a close, and, before we hop on a flight back to the UK, we spent an hour or two typing up just a few snippets from some of the interviews we conducted at the conference. So, if you'd like to read what Jacob Kaplan-Moss thinks about Google, what Jim Zemlin thinks about Larry Lessig, what Michael Tiemann thinks about lobbyists, what Stormy Peters thinks about KDE, what Evan Prodromou thinks about Miguel de Icaza and what Bradley Kuhn thinks about Mono, read on...
These represent a small portion (around 500 words each) of some of our interviews. We talked to far more people and everyone said far more than we have space for here, but we're typing them up as fast as we can and will get them into Linux Format magazine as soon as we can. If there are typos below, they are obviously entirely our fault.
(Jacob was at OSCON to talk about Django, the open source web application framework built on Python. His Mac broke straight after his talk, so he installed stacks of apps on his iPhone so he could stay on the #OSCON IRC channel.)
Would you say that web frameworks are a zero-sum game?
You know, in the early days, I was really obsessed with "we gotta win, we gotta be bigger, we gotta have more users, we gotta be better, we gotta be awesomer!" And I can't really put my finger on it, but at some point I just kinda said, "why?" If you're a Rails user and I convince you to use Django, what does that do for either of us? So lately I really haven't done any evangelism per se. My tone has changed from "you should use Django because..." to "I use Django because..." And I feel that I've started to get pretty unhappy with the tone of competitiveness within open source communities. "Oh, you use Vi? I must kill you!" That is a zero-sum game, but the real battle as I see it is between closed and open. As long as we're all using open source, we're all winning.
It must have been a pretty good PR boost when Google announced it was using Django in AppEngine. But I think it was 0.9.something that they went with...
Yeah; Google doesn't co-ordinate with anyone on their release process. They just release stuff, and, you know... frankly if someone from Google had said, "you have to release 1.0 now because we want to get AppEngine out, we'd have said "go stuff yourself." I think that's necessarily a problem, but, yeah, it was pretty cool. I found out about it when everyone else did. I was kinda surprised that they didn't reach out at all to anyone in our community to try to have any sort of discussion or integration or anything, but I guess that's the Google way - they don't talk about upcoming products.
But you've suddenly gained a much larger audience.
Sort of... it's more like using a fork of Django than actually using Django, because the model layer is completely different. Anything built on Django that needs that model layer, that includes the admin interface, that includes the authorisation framework, and that includes a bunch of other things, just doesn't work on AppEngine. And I we can do some stuff to solve that technically, that mismatch, but there's always going to be a impedance because Django basically assumes a relational database, or at least the call parts do, so you're never going to be able to take some app you've deployed on Postgres and deploy it on AppEngine, because the way you have to think is completely different.
Yeah, Google have said, "you can have Big Table or... nothing."
Right. In the same way they said, "you have any form of authentication you want, as long as it's Google."
And I think they have a strange version of SQL...
Oh, I hate that actually. GQL? Something like that. So we've got GQL, the Google Query Language, and we've got FQL, the Facebook Query Language. Great, great... wonderful. Because SQL was too standardised.
Presumably you now have all these new users who are using an old version of Django, though...
No, they bumped it up to 1.0 now. And I've been told that once we get off our asses and release 1.1, they'll get off their asses and put 1.1 in AppEngine. I think the reason it took them a while to get 1.0 out was that they did the whole Release Early thing, and one of the side effects of that is they didn't have a mechanism to have multiple versions of frameworks available. I think they do now.
So is Django 1.0 thread-safe now?
Well... <taps the table>.... <silence>... Well I run Django in production in threaded mode, and it works. Which is one definition of thread-safe! You know, threading bugs are the worst sort of bugs. We went through a lot of effort to make sure the parts we knew were unsafe became safe. So we're into the unknown unknown territory now!
(Jim is the executive director of the Linux Foundation. He makes awesome hand gestures on the table as he talks, but probably doesn't realise.)
So... obviously very high on your agenda right now is Open Source For America.
Yeah, because we just announced that... today? Right, today.
How involved are you in that?
I'm involved. I'm not central in running the organisation or anything like that. This is something that people like Tom Raybon over at Red Hat, or the folks are Sun are reading more than I am. But I do think it's a good time right now, with the change of administration and the mood in Washington to show the federal government how they can benefit from open source. I think the key thing to understand about the mood in DC today is that the country is demanding lower-cost of government and more accountability and transparency in government, and that's something that open source, with its open model, exemplifies. So in many ways the timing is really perfect here. This organisation is going to work with members of congress, members of the administration to show them how they can benefit from using open source to control government costs.
Would you say that many geeks have a natural distrust of lobbyists?
So... a typical definition of a lobbyist is someone who raises campaign funds, contributes to members of congress or political candidates campaigns, and that's largely understood to give them access to make their point of view to candidates. This organisation is not engaged in that activity right now.
It's not a political fund-raising organisation, it's not making campaign contributions, really this is an organisation that is an advocacy group to educate the federal government on how to benefit from open source.
Well, there's still money coming in for the purpose of bending ears in government...
There's no money in the organisation - it's largely made of volunteers through contributions of people's time through the member companies of the organisation.
<an iPhone beeps>
Is that us?
It's me. Go ahead... you were just asking...
No, no; we can wait.
I'm just going to turn it off...
So... you've got an iPhone? For the record?
I have a dozen phones.
Do you have an iPhone?
I've got an iPhone. I've got a Palm Pre. I've got a Google G1. I've got...
On you? Right now?
Not on me right now...
So the phone you're carrying right now is an iPhone?
Yeah - gotcha! (laughs)
(Disclaimer: TuxRadar podcast listeners will know we have no problem with iPhones and that several of our team own one)
So to get back to the group, though... these organisations are made up of a committee of people who work collaboratively, there's no money involved.
Well, there is money somewhere; someone has to hire these people to do the volunteer work...
Sure, they are full-time employees of the member organisations, who give their time...
Their work time or their spare time?
Their work time, so if you could allocate a dollar value to that... but again to be clear: a lobbyist, in the traditional definition, registers to engage in the activity of fund-raising.
Right, but my definition in my head is people who are in politics trying to talk to decision makers, saying "hey, have you considered open source?" again and again. Is that basically what the plan is? Is that a fair description?
Yeah, yeah - definitely.
Well, certainly I think a lot of geeks have a distrust of lobbying in that definition. Larry Lessig is definitely against the idea of lobbying politics--
But everyone likes him!
Everyone does like him, yeah. He's likeable!
So there's good lobbyists and there's bad lobbyists. We'd like to consider ourselves the Larry Lessigs, not the Jack Abramoffs.
So, competition is very good, but you have Linux.com doing news, you have Distribution Central providing distro news and updates, there are forums, there are events, there's training... do you not feel you're kind of muscling the community out of the way a bit by doing it through the Linux Foundation?
It's a community-generated site. There are 7000+ users, there are 400+ users, there are...
Sure, but what I'm saying is that before the Linux Foundation came along, we had lots of forums doing very well for themselves - distro-specific stuff, such as Ubuntu Forums or OpenSUSE's forums... for conferences there are things like OSCON, or Linux.conf.au... all doing very well...
We're muscling Linux Questions out of the market...?
Well, I'm asking you!
There's a quote from your website here. "Why receive training from anyone other than the Linux Foundation?"
Where is that?
That actually scared me a little.
Because you're saying not to go anywhere else! The problem is that you have certain people working inside the foundation, which adds a huge amount of weight. You know, if the people who employ Torvalds say "don't go anywhere else; come here", that's scary for me.
Retract that immediately. It's absurd. There are lots of places to get Linux training.
(NB: Jim was really good about this - he got on the phone immediately, during the interview, to get the statement removed from the site, and it was indeed retracted a few hours later.)
Linux Foundation website: Before. "Why receive training from anyone other than the Linux Foundation?"
Linux Foundation website: After.
(Michael is known for his incredible work on GCC and GDB, but he's also the VP for Open Source Affairs at Red Hat. He was at OSCON representing the OSI, which he is president of. He's warm, funny and always happy to be straight with his views.)
So... you must be really jazzed right now.
It's a really exciting time right now... the announcement this morning...
Absolutely. Open Source for America is bringing together 60+ companies and organisations to act as a clearing house for information, but also to be able to take specific recommendations to people in power to implement them. The OSI is a little bit different. As a US non-profit we try really really hard not to cross over that threshold that could be interpreted as taking any kind of lobbying activity per se. Our charter is education and advocacy. And we're delighted that Open Source for America consulted with us, brought me on to their advisory board, and also has adopted.. I can't tell you how many national policies I've seen, but certainly in their neighbourhood of more than a dozen, either draft or enacted, that specifically reference the Open Source Definition from the OSI as being the defining document of what must a licence be to be open source. And moreover Open Source for America also recognises our standard, which is that only software covered by an OSI-approved licence is open source.
Well, I was speaking to Jim Zemlin earlier, and I said that some geeks are naturally distrustful of lobbying in the government. And he said "no, we're not lobbyists; lobbying is when you give money to campaigns" That's not how I think of it. I think of lobbying as bending ears in governmental places...
I'm with you. I may not have the precise definition, but in my view, if you sit in someone's lobby and wait to tell them "hey, I've got a good idea, you should really look at XYZ." I have no idea whether money influences that, but I do know that it is possible to publicly access one's elected officials, and to the extent that one is involved in a specific recommendation, I would call that lobbying. But that shows what I know!
I think so too. But I think the problem is that we're both geeks, and if we think that then so might other geeks. And if geeks are distrustful of lobbying...
I think the issue there is that you have to look at how the overall world works. I certainly do know people that are so cynical about the act of voting that they choose not to vote, right? I choose to vote, and then I pray that my vote is counted.
So do you think geeks fear lobbyists? Or should fear lobbyists?
I think there is a prevalent mistrust of government among today's geeks. At the session I just came from, Open Source, Open Government, I looked out into the audience and I recognised many of the faces there as being unreformed libertarians, who take as one of their central tenets that most government is just bad. And what I admonished people to do as open source becomes more popular in government is to recognise that rather than reform the libertarians in their understanding of government, we should really bring the message of open source to those who are much more actively engaged and participating.
It sounds like what you're saying, and please correct me if I'm wrong, is "all those lobbyists are bad, but don't worry: we'll be different, we'll be good lobbyists." That's rather like a certain Linux company that says, "patents are bad, but we'll register patents as well."
Well, I guess what I would say is that I'm a big believer in trying to meet people where they are. And so if somebody's fundamental attitude is to be fundamentally anti-government, then I'm not going to spend a lot of time trying to talk to them about how to work with government. Or to try to explain why a particular approach to government is magically good. But what I can definitely tell you is that there's a huge population of people who are civically minded and civically engaged who don't know what they need to know about open source. So I guess the answer to that question is, again, the role of the OSI is not so much to minister to those who have already convinced themselves of all the benefits and value of open source. The role of the OSI is to expand the population of those who are aware of all those things, and who can successfully promote them.
(Stormy is executive director of the Gnome Foundation. She smiles a lot and really means it.)
Now that you're working on Gnome 3.0, How much has the Gnome Foundation learned from KDE 4?
I hope that we've learned a lot. KDE is our sister project on the free desktop; it's our competitor on the free desktop...
It's a competitor?
I always say it's like a competitor like two players on a team competing for a Player of the Year award. They decide to do things a little differently than us, so it's interesting to see how that goes. So we watched with great interest when 4.0 was coming out, to see what was happening. They did some good things, and they did some things they probably wouldn't do again, and we hope we learned from those.
Well, certainly one of the things the KDE 4.0 guys said, and more recently also the KOffice 2.0 guys said, was that "hey, this release, this major version number release, it's not for end users."
Yeah, I don't think we'll do that. But I do think the distributions might do that. So they might say, "Gnome 3? We just haven't had enough time to play with it, so we're not going to put into the very next release we put out."
The thing everyone is talking about is talking about is Gnome Shell, largely because it's very whooshy - it's very impressive. How important is that?
I just talked to a very large financial company that wants to move to the Linux desktop, to move to Gnome. He said he needs those whizzy things. He's trying to talk people into using Linux, people who've never used Linux, he says it helps a little bit to add some whizzy stuff. Like iPhones, right? They look nice. He needs it to be functional, but he needs a bit of whizzy to help sell it.
That's one thing Gnome has always done well - it looks polished, it looks refined.
I started this job last year, and the thing that struck me was how much people kept talking about how beautiful things are, and how pretty they are. It's definitely a value.
But do you not think that, thanks to Canonical, most of the world thinks Gnome is brown?
We just started launching our Gnome Marketing Effort.
I saw you mentioned a few months ago the idea of having some sort of logo, "Gnome Inside".
Yeah, so we've talking about that. And we've been talking about who's our audience for marketing. Do we just market to distributions that ship our product? Do we market to end users? And if we market to the end users, how do we engage the distributions in that marketing? So I think it's a work in progress right now.
Well, clearly Mozilla hit some problems when they forced their branding too strongly, and now we have Iceweasel.
Right, so Gnome's goal is to have a desktop accessible to everybody. So I think we want to work with distributions to make sure that we can bring a desktop to everybody and not fight about brands in the process.
But you want the Gnome brand in there somewhere?
We're working out how and what, but yes. You can still do an About Gnome right now and see the window.
Warning: contains strong(er) language
(Evan is one of the smart geeks behind Identi.ca. He sticks out at OSCON because it doesn't look like he dresses in the dark.)
Very early on, you sent some success metrics. What I thought happened was that a lot of geeks worked hard to promote Identi.ca early on. On the flip side, though, perhaps the platform wasn't quite as developed as it might have been. Here's what Miguel said, "How do I find people on this thing? Why does this thing not support my font settings? THIS IS HORRIBLE." He left, and hasn't come back.
This was day one, right?
It was, and he hasn't come back. Do you think in a way the initial push may have worked against you by making people think it's not quite ready yet and going back to Twitter.
Right, so what really happened there was that Miguel de Icaza is totally useless, someone who was totally not a big supporter of us ever. He was a total dick! I mean, pardon my French, but he was a total dick. I mean, he came onto the site. This guy... the guy who started Gnome? Gnome was a piece of shit for, like, many years, and he comes onto our site and he can't figure out how to use the buttons, and he says all that nasty stuff? I've got nothing for him. He's a jerk. He was a total jerk about it. So what happened with Identi.ca is we had about 150 people when we were in a private beta time. And then I told people, start telling other people. And we got swamped.
We got a lot of people who had heard about it and had no idea what was going on, and they came, and they said "this is the next Twitter; let's do this." And it was version 0.4 of the software. We're still at 0.8. I really tell people "we're still beta software, this is what we're doing; you are seeing the development of the software, you're not seeing the launch of a finished product." And so people had different expectations, and that's why someone like Miguel would come in and say something like that. But also because he's a jerk.
Do you think, then, that in the long term people will start coming back and saying, "hey, these features are here now, I can use Identi.ca again"?
Absolutely - we get that all the time. People went back to Gnome, right? And there were people who stuck with it, like me, and never said nasty things about it. But no, people have been coming back a lot. And one of the things that has really happened is that we've had a lot of innovation in the last year. We've added a lot of things people expected from Twitter, like replies, like tagging.
Some of the things that happen in the greater ecosphere in Twitter happen now for us, within our core engine, which is nice because we think they are core features. So things like file sharing, images, groups, and subscription management. All that says the platform is still evolving. And everything we get from people who have gone away and come back is, "wow, this is looking really good."
So do you think, as Twitter adds new features, you're chasing a moving target?
I think we're doing different stuff. We're not trying to tail their features in the same way. I think we are providing two different tools that do the same thing. And the user experience of Twitter is different to the user experience of Identi.ca. And they have certain constraints, and so do we.
Bradley is technology director at the Software Freedom Law Center. He's also an incredible poker player and arguably one of the fastest-talking people we've ever met - this entire response lasted a minute or so!
What exactly is your stance on Microsoft's Community Promise for Mono?
So here's the thing. I want to trace the history a little so people really understand what's going on. So what's happened is this language has become popular: C#. And it's infrastructure, CLR, CLI. It's being implemented as Free Software. And I want to say first and foremost: it is a good thing to get that stuff implemented as Free Software. I think Miguel is running an important project. Because there's a real reason we need it: people who are living in a Microsoft world, and working in C#, .NET, the CLR, etc, when they come over to the Free Software world and install Ubuntu or Fedora, they should have the same development tools they are comfortable with. So for that reason alone I think it's really important that we have a full CLR/CLI implementation working just like it should.
Now the question, and the reason the debate started, is whether or not we should let that technology be the default technology for when we're building new, important Free Software projects. And that's the only place anybody is debating - no one's saying the project isn't valuable, because it definitely is. But when you want to write a new, important application, what language system should you pick? What I've being arguing for is that we should pick an infrastructure that mitigates our risk as Free Software developers, and, it turns out that Microsoft is well known to be using patents against Free Sofware. They did that in the TomTom case. 18 months ago Steve Ballmer announced "I'm going to attack Linux with patents." And then we see the TomTom case.
Well, the TomTom case is the tip of the iceberg, because, as I've talked about earlier, it's rare things make it all the way to litigation. It's the tip of the iceberg: we know they are out there enforcing against other people who settled out of court. So they are going around shaking down people for patent licences on the FAT patents, on all sorts of other patents that read on Free Software. We know that's happening - it's out there, and you see it. And just recently, just last week, they again went after a company, Buffalo. They enforced patents against them. So we know Microsoft has a strategy to enforce patents against Free Software.
Usually they go after companies, because the other point I want to make is that individual developers don't have a lot to fear, because Microsoft's not going to sue an individual developer because they don't have a lot of money to go after. What they'll do is that the individual developer writes something really cool, releases it under a Free Software licence, a company picks that up, and then Microsoft goes after the company. So when developers are making these decisions, it's not that they have to be afraid for themselves, and I really actually generally encourage developers not to sit in fear in patents. You should do your work and write interesting software and not sit in fear of patents. But you should think about what the risks might be with regards to broad strokes questions, like which language to use.
So because Microsoft has this stance against Free Software using patents, we have to be particularly afraid of them. And when we compare it against languages from our community, like Python, Perl, PHP... technologies we developed as a Free Software community, the patent risk is somewhat mitigated. It's not gone. But it means there's not one company with a master plan to come after Free Software, which is Microsoft's master plan, make no mistake.
Some years ago, the OSI announced it was going to be running awards for Open Source, cunningly called the Open Source Awards. The largest award, for "Grand Masters", was to be given $10,000 prize money, and the winner, announced during a big event at an earlier OSCON, was Larry Wall for his Patch program.
A couple of years later, I asked Larry whether he had received the money, and he said he hadn't. This year I spoke to Allison Randall about the money, and she said:
Allison: "You know, I never heard back about that. I think he did eventually. I'd have to check with Larry. He certainly doesn't complain about it if he didn't."
Us: "He's not the kind of man that would."
Allison: "I know, he's not. I think I remember that he did get it, but it was like two years later. That's how I think the story went, but I'd have to go back and check through my emails to be sure."
So, the next time I wandered past Larry in the hallways (and you do - that's the nice thing about OSCON), I asked him and his wife Gloria whether they had received the money. Here's what they said:
Us: "Did you ever receive the $10,000?"
Larry: "No. I've kinda written that one off."
Gloria: "Well, maybe you did receive it as part of another payment. You might not have noticed."
Us: "Surely you'd have noticed $10,000?"
Gloria: "Oh, $10,000! I thought you said $1,000! Yeah, we'd have noticed that alright!"
Us: "Larry, there's humble and there's humble. You need to say something."
Larry just shrugged and smiled. Anyone who knows him knows that he's that kind of guy - he doesn't like to complain, he doesn't like to make a fuss over things, he doesn't like to bother other people, and even though he has a large family and was unemployed for years he still didn't get upset about not being paid. But without the fantastic work he has done for the Free Software community - large chunks of which were done for no payment - we'd all be much poorer.
So, OSCON 2009 is now drawing to a close, but I think this issue needs to be settled once and for all: Eric Raymond gave the award on behalf of the OSI at an O'Reilly conference. Surely someone must know where the money is? Please, readers, don't just leave this page and think "there were some neat quotes in that piece..." Do something. Say something. Help get Larry his money.