Jack Aboutboul - interview

From LXF Wiki

The Personal Touch

Imagine that you could design your own distro online, just downloading an ISO containing the packages you needed... With Revisor, you can. Fedora's Jack Aboutboul speaks.

The day this very issue of Linux Format goes on sale, Fedora 8 will be entering its last development freeze, with a full final distro release on 8 November 2007. A couple of months earlier, Nick Veitch caught up with the "Godfather" of Fedora ambassadors, Jack Aboutboul, to talk about the upcoming distro release and the future of Fedora.

Jack started out as a Linux user a decade ago, serving as an early developer on Gentoo Linux and was a founder member of Syracuse University LUG. Also, as primary developer of EduGate, an educational institution resource management suite, he recently founded the Jack Aboutboul Foundation in late 2007 with the mission of "building bridges between open technology, open content, art and education, in a directed effort to enrich the lives of people and children worldwide; and to animate the creative drive which is present within each human being." http://jackfoundation.com.

Linux Format: Explain yourself, what's all this "Godfather" stuff about, then? Have you been caught putting horse-heads in people's beds?

Jack Aboutboul: I have Greg DeKoenigsberg, Red Hat's Community Relations Director, to thank for that grandiose honorific! My official title is Community Engineer at Red Hat, where I focus most of my efforts on the Fedora Project. In the past, I've helped found some of the Fedora Community initiatives, such as the Fedora BugZappers Triage Team; I created and help run FUDCon, the Fedora Users and Developer Conference http://fedoraproject.org/wiki/FUDCon. I'm still doing pretty much the same stuff as I've always done, working on the interesting projects.

LXF: What sort of interesting projects?

JA: Over this summer, we've been working on Revisor, http://revisor.fedoraunity.org, the regraphical Respin tool; and also we've been working that into a web service so that basically we can log on from anywhere essentially and create your own custom spin and it'll go out and build it for you, email you when you have an ISO ready to download. The credit goes to the interns for all their hard work. Revisor is really cool because a lot of what people complain about with distros in general is how massive they are and that they don't want to download the whole thing when all they need is a few packages. Some people really do want it to run on just a router or similar; with Revisor, you can go online and put together a 50MB ISO for instance, and it'll ping them when it's done and they can download it. That has applications for different things, not just for people with business implications, so that's been kinda cool. We've been working on that pretty full force and that just went into public beta last week. (For more on Revisor, see this month's cover feature.)

LXF: For individuals wanting to do an individual project that's really useful, but won't companies that use Red Hat want to use it as well?

JA: Of course! The eventual aim is to be able to offer support to some of these respun distros. Red Hat is committed to the new software paradigm in the business world. "To be the defining technology company of the 21st century" is written in big letters in the reception. The first time I saw that I thought "you're kidding!" but as time goes on, I find that aim less and less funny ­ it's a serious proposition.

It's really good to be able to put together a software appliance that makes it really easy to just be able to go out and customise the distro to suit individual users' own needs. With the tools we provide, it becomes so much more easy to put together something you need for yourself and then use it like that; the reasoning behind that is that individuals have a better grasp of what it is they need than we do, obviously. The business implications of what we're doing online is we want to be able to get to a point where we can create hosted stacks. We're working with various web service companies on the project to, so we want it to be able to do something like spit out an image for you ­ you create a subset of Fedora with some extra packages of your own and hopefully you eventually plan on getting those into Fedora as well and contributing back. You know the ideal is not always what happens so being able to spin that and pushing that out as an ISO to some sort of hosting space and then being able to run that from there. Whether Red Hat would like to extend that to sell support to something like that or sell an offering to something like that, well it's not in my hands, but it does become possible, the possibility exists. And I don't think we're not going to take advantage of any possibility that exists.

LXF: I guess that in the past, Red Hat and Fedora have been considered to be one of the harder distros to customise. I suppose that's why some people may have looked elsewhere in terms of personalising their distro for their own particular needs, and now with the developments that you've been working on, that's not the case anymore?

JA: Well to tell you the truth, everything seems really hard to customise! I've tried playing with other distributions and to try and see, how easy or difficult it is to do a respin type of set up to see what you can pull together. I mean, Klaus Knopper is awesome: I don't know how he does it because it's just a tremendous amount of work to get something like that going. Knoppix is a great distro, and mostly the work of one individual; just the fact that he was able to pull that off was a feat in itself. And then you have a couple of months later everyone trying to catch up and standardise how this whole process actually happens, and that was part of the problem with us trying to that is that we had two conflicting sets of live image creation tools and so we started out with something called Pilgrim. Then we started writing something completely new which at the time was also called Live CD Tools and we were on the horns of a dilemma: if we decided to go with this one then these people are gonna get mad, but if we go with that one then those people are going to get mad, so what do we do? And really neither of those apps fitted the bill of what we needed to do completely. That's how Revisor came about. It was just people from the community saying "Hey, we're writing a new tool." At FUDcon, we invited people and paid for people to come over, and we thought "Why don't we use Revisor as our project for the weekend?" Right there, we went from something not existing to having something half-workable; and obviously it's been getting better from there. This is a brilliant example of how effective it can be to open up and let the community contribute.

That was also a problem that we had before ­ last year when I met LXF, we were talking about the intention to make Fedora completely open. That was a major hurdle: some people within Red Hat who thought that this would be a miserable idea and that all hell would break loose: that we'd lose control of the things that we needed to control. There were Fedora developers that were overawed by the scale of the undertaking that was needed. There was a tremendous amount of work that needed to be done. We needed to make sure that IP concerns were taken care of ­ we really broke our backs to do it.

LXF: I remember that there was a lot of Fedora packages that had incorrect or outdated licensing information. Going through those packages to check all that information seems like it was a Herculean task in itself...

JA: It was all done by hand! We started that at FUDcon too, We set up 20 machines in the lab, people sat down and started reviewing licences, checking packages and making sure that things fit the standards that we defined for our packaging guidelines. It's intensive manual labour, but we did it! I don't think people realise the enormity of the task unless they've been involved with a big project like a distro themselves. We had an announcement the day we did it, and a lot of people weren't sure what we'd actually achieved!

The people that were most grateful were folks from other software companies who had been building appliances off our software ­ we'd just made their lives a whole lot easier. Completing this really fostered the sense of community between ourselves and the people that are dependent on us doing our jobs right. Creative Commons was a major factor that we hadn't really considered that much up to this point, but it really helped!

LXF: I was talking to Jon Phillips about how Linux Format could help with the documentation side of things for the Linux community as a whole. We have a huge back catalogue of articles and images that we are trying to work out how to release without damaging our circulation figures...

JA: The benefits in the long run will be boundless ­ including LXF tutorials in distros would be a neat way to get your words out and about! I believe that from the very earliest implementations of the Fedora distro, we have wanted to set ourselves up to be a platform for innovation. Whether we innovate ourselves, or can enable and encourage others to do so, that's important. Putting the tools into people's hands is what we do.

There's probably two main avenues from which innovation is going to emerge. Obviously, there's the technical innovation, which we've been on top of; and that's what we're known for, is our technical excellence, if not anything else. We're going to stick to that: we're not going to throw our core competency out the window for no good reason.

Then there's the atmosphere of learning that not just Red Hat encourages, but all Linux projects ­ there's this rarefied atmosphere of freedom which encourages people of all levels to develop solutions to problems without being tied into a particular way of thinking, limited by corporate guidelines and restrictive patents.

LXF: Your custom distro idea is almost like inventing a brand new media format. Being able to remix the distro to deliver what each user or group of users want is almost like opening a new TV channel...

JA: You don't know where the next great idea is going to come from. Instead of going into the computing world and spending millions of dollars on R&D, or buying up companies who look like they have some interesting IP or crazy hack projects that never get anywhere that you can fold into your own business model. Why do that? It's just a matter of empowering the community. We can sit back and relax, and the great ideas make themselves known to you through the community's process of peer review.

Michael Tiemann, Vice President of Open Source Affairs at Red Hat gives an enlightening talk, The Open Source Triple Play [www.redhat.com/magazine/001nov04/features/tripleplay/] describes this situation perfectly. It's users who decide the direction of any sort of software project, whether in the proprietary sphere or not. Why bother trying to steal their ideas and implement them, when they're capable of contributing to the project on an equal footing, that saves you money into the bargain?

LXF: That's always a classic problem in software engineering anywhere. Talking about the Fedora community now, it was still fairly unformed when we last spoke in 2006, still mainly being driven by Red Hat people. The community itself has grown a lot since then, as shown by the activity on mailing lists, developers' wikis, blogs etc.

JA: It has grown a tremendous amount, thanks are due in a large part to the way that the governance model was set up. The way in which we let smaller interest groups have their say is inclusive, but without being disruptive to the main direction of the project's big picture. The readoption of all the KDE parts of Fedora was something that worried us to start with, but the community seemed to eat it up: they just got on with the task in hand. By offering support to those groups making the most headway, we could streamline everything. Of course, we're helped also by having some KDE developers sitting on the board now! It has worked out really well. We have a KDE Live CD now, everything is up to date. We're impressed with recent developments in Gnome too: the acronym for Gnome Online Desktop ­ GOD ­ I love it!

LXF: There might be some complaints about that! Personally, I use KDE on my main work machine in the magazine office. So, the propitious question to ask: will Fedora 8 be shipping on time? Or have you decided to move the date back to the New Year so you can incorporate KDE 4?

JA: We're going ahead with the early November release, whether KDE 4 arrives or not. We've been pretty on-schedule so far. For all releases except Fedora 6, we've made our targets. From now on, we're aiming for May Day and Halloween releases as this seems to line up with the release schedules of everyone else; it's better to have a rolling schedule.

We've been doing a lot of QA recently. Will Woods, who's heading up Fedora QA has been doing a fantastic job. Even with Fedora 7 we had a few small things slip under the radar, so we want to limit that happening in the future.

LXF: They did get fixed pretty quickly though. With some distros, you install on the day of release and discover a week later that there are 200 packages to update. There was a noticeable step forward though: 7 was less broken than 6 was.

JA: Fedora 6 was actually a pretty good release, all things considered! We're never happy with status quo, we're always trying to cover new ground with the testing and making sure that people have access to it earlier and that there are more people in the loop.

LXF: What do you think about Dell's adoption of Ubuntu for pre-loading on its hardware? Do you think that Fedora should be aiming to get more prominent in that sort of area in 2008?

JA: One part of me obviously says "Yes!" but the smarter part of me says "wait" What Ubuntu has been doing is great, as we've all been trying to get Linux as a preinstall option for computer buyers for years. Obviously we disagree with some of the minutiae like non-open codecs and drivers, but their experiences will be shared with us so we can go about solving those issues in what we see to be the correct way. We're working on things like Smolt, the hardware profiler. We want to be able to gather statistics and approach companies that produce things like wireless cards, and make a convincing case for them to open their drivers based on the number of users that a distro like Fedora has. It's two-way ­ if Fedora users discover a bug in the wireless companies' drivers for instance, we can alert the company to the fact and even help them fix it.

LXF: I never understand why companies insist on closed source drivers. If a company made something like a PCMCIA adaptor, that supported Linux, everyone that ran Linux would buy that piece of hardware, as it `just works'.

JA: A common excuse used by hardware companies is that "we don't have the engineering resources to support that," but this is a fallacious argument, as they don't need the resources ­ just give people the tools and they will do it for you. It shows, sadly, that a lot of businesses in the computer sphere still don't have the remotest understanding about the ideals of open source. The IP issue is something that scares them. Sun is experiencing this problem right now, as one of its projects wants to use some video drivers that are closed source from companies that don't even want to entertain the idea of benefiting from opening their code. It's a chicken-and-egg problem. Hopefully through things like Fedora's Smolt hardware profiler, they'll get clued up and what they see as the political aspect of open source will become less of a problem when they see the benefits to their bottom-line.

AMD announced at the Red Hat summit that they're prepared to open some of their drivers [as reported in News this issue, page 11] The Fedora community has offered to help, and they're seeming very receptive. "Doing things the right way" has always been the Fedora mantra, and that gives us a bit of a personality problem when it comes to Ubuntu. People load that distro up and though it `just works' they're still no wiser about the issues involved in how the wider Linux community can go about changing the closed source situation. I don't think it's doing harm, but I personally think that it's encouraging newer people to the Linux world to take things for granted...

LXF: The user base for Linux has changed greatly in the last two or three years though. In the past it was a conscious decision: often political or ideological: users bought into the concept of free software; now, in some ways, because it's popular, some users don't care and just favour Linux for purely technical or convenience reasons. An example would be the rise in popularity that Linux seems to have in the home studio/electronic music market ­ some of LXF's sister magazines carry Ubuntu on their discs purely as a platform for their readers to be able to use open source audio production applications that aren't available for Windows or OS X.

JA: Fedora has been working on the Mugshot platform for two years now. The online desktop is the next development in people's computing habits. Everyone's using Gmail, Facebook, RSS, all these online services; that's a consideration that distros like Ubuntu have yet to identify, so in purely online desktop terms, Ubuntu is lagging behind Fedora in this area. We're not trying to compete, nor are we just innovating for its own sake: by different distros concentrating on different areas, the choice that will be offered in the long term by Linux as a whole will completely outpace the proprietary offerings.

I can't speak for Havoc Pennington (maintainer of the Mugshot social networking website, http://mugshot.org, that aggregates content from MySpace, YouTube, Facebook, Flickr, blogs and more all in one place) but I am convinced that we've really outdone ourselves with Mugshot: we're at the forefront, nobody's even close to touching us in that space.

Fedora's dedication to opening everything is not just for hackers ­ it has a wider importance in that our approach is an agent for social change. That's the reason that we love to work with Creative Commons, as it pushes copyright reform: and the changes that are needed there will effect everyone, way beyond those of us who like to hack our computers. Old school hackers like Brian Behlendorf (Chief Technology Officer, CollabNet) are always looking for the next big thing, so we're hoping to be at the forefront of whatever's happening, in order to help things be "done right." We want to provide a solution for as broad a number of users as possible; people with particular needs can split their own distro.

LXF: We have the same issue trying to identify who the readers are of our magazine. Are they IT pros? Users? Developers? In the end we realised that it didn't matter hugely, what we do is explain things and show them how to do things. It doesn't matter whether they use it for business, or whether they're just doing it because they're interested in it. It's important that both paper publications and Linux distros don't try to over-think the situation and pigeonhole their audience.

JA: Each person is unique, and so is each business. The uniqueness of the free software movement can better serve people's needs by being more responsive to those considerations than a locked-in proprietary offering that `almost fits' Although choice is important, the open source movement needs to unify in such a way that there isn't too much duplication of effort, though reinventing the wheel is sometimes necessary ­ to get better wheels,

LXF: What big challenges are there for Fedora in the next year?

JA: Identity is perhaps the biggest. What is Fedora? We're kind of in a grey area right now. Are we a distro? Are we a platform for a bunch of other things? Are we a system by which to create software appliances?

LXF: What's the answer?

JA: Many things to many people. The battleground for the next year is one of differentiation. How do we stand apart from the pack? I recently read a book called How Soccer Explains The World (HarperCollins, ISBN 0-066-21234-0); in it, a theme that Franklin Foer identifies is that we've come to a point where it's commodified tribalism: it's so easy to slip in and out. It's becoming the same with Linux: One week you use Fedora, one week you use Ubuntu, one week you use something else. It's all the same, it looks for the config files in the same place, I look for my settings and they're all there... People want more than that; I think it's our duty to offer them that. There's some interesting stuff in the pipeline that will redefine the Fedora brand, and identity, and maybe retake the top spot. I'm not worried: when Gentoo came out, everyone enthused about that. The pattern has been repeated often. LXF