Interviewee: Adam Williamson
In this interview we talk with Adam. In specific, we talk about:
- What’s new in the latest Mandriva release
- Relating directly to the users instead of the developer community
- Differences between user priorities and developer priorities
- Deciding what should be included in the distribution
- What will become of the “year of the Linux desktop” idea?
- The future of open versus closed development
Sean Campbell: Adam, please give us some background on you and on Mandriva.
Adam Williamson: Sure. I’m the community manager for Mandriva, which is my most important role with the project, but like everyone else here, I do about twenty other things as well.
I’m also the bug tester, I maintain several packages, and I do a lot of public relations community announcement stuff. Mandriva is one of the leading Linux distributions. It’s been around since 1998, so it’s one of the older ones. It primarily has a desktop focus, and we’ve got a strong international community.
Sean: Since you just released the most recent version, walk us through the most significant features you added to that release.
Adam: The biggest new feature is that we are using KDE 4 as the default desktop. We had a sort of testing version of it in previous releases, but it was more for playing around with than really using. For this version, KDE 4 is the default, so when users install, that’s what they will be using, out of the box.
That change brings a lot of great new features with it, like the plasmoids, the new file manager, Phonon, and all that good stuff. We’ve also overhauled our tools and our install. We’ve given them a completely new appearance, so they look nice and it also improves the ergonomics a bit.
Aside from that, there’s a whole laundry list of stuff. We’ve got GNOME 2.24, as well as LXDE, which is good for netbooks. We have really good netbook support in this release–we test things on the Eee PC, and we also have it working on the Aspire and the Wind. We’ve improved the boot speed quite a lot.
Something that I personally am involved in is support for working with mobile devices: Windows Mobile, BlackBerrys, and things like that, and I’m pretty sure we have the best support for that out of any distribution.
Sean: I remember from our previous conversation that you guys have invested a lot of time in mobile device support. Looking back historically, what particular challenges have you overcome in trying to integrate mobile device support?
Adam: That’s actually my stuff. It started with the 2008 Spring release. One of the nice things about Mandriva is that it’s a very flexible system to work with. About two weeks before 2008 Spring came out, my partner–who works for a cell phone store–brought home a Windows mobile phone. I plugged it in and couldn’t do anything with it. I couldn’t even look at what was on there, let alone start synchronizing it, so I decided to make it work.
There’s a framework called SynCE for Windows mobile devices that plugs into a thing called OpenSync, and then you need an OpenSync front end to actually synchronize anything. It’s a very complicated process, and at the time of 2008 Spring it was even messier because lots of this stuff was broken.
It involved getting OpenSync, SynCE, and a front end called Kitchen Sync all packaged up properly, and patched and working together. It was a matter of little bits of talking to various projects and getting patches and working out a couple of things that I had to fix myself.
I also had a BlackBerry lying around, so I got on a bit of a roll and I thought, well, I’ll get BlackBerrys working as well. It’s one of those little inspiration challenges, and the interesting thing is that it’s something that probably real people do more than distribution people. Your average distro guy probably doesn’t use something like a BlackBerry or a Windows Mobile phone; it’s not his work system.
But in my role of community manager, I know there are quite a lot of people out there who would find this quite useful. It’s kind of something that falls between the cracks for a lot of distribution developers, because they just don’t work that way.
Sean: Given that you’re the community manager, what are the features in Mandriva that you feel the development community rallies around–and you don’t really have to work hard to get them to throw their effort in–and what are the features that you have to sometimes say, “Hey, I need you guys to help out over here?”
Adam: In most companies, the community manager talks mostly with the development community, but I actually work mostly with the user community.
Our development is very collaborative between the community and the internal developers. It’s all done in Cooker, which is the open development distro; we don’t do anything behind closed doors. We don’t really have to drive anyone to work on any particular area, because we find that people tend to work on areas that are useful.
We don’t have a community manager in the sense of someone who directs the development community and says please do this, please do that. We’ve been doing community development for a long time, practically since we started, and we’ve got a lot of long-term contributors who understand what needs to be done. It’s really not something that we’ve had a problem with.
Sean: Given that you’re more focused on the user community, what kinds of things has the user community asked for that you realize are going to take longer to implement than the community thinks? This issue seems to come up a lot, when users are calling for a specific feature, but since they’re not software developers, they may not really understand that what seems like a simple request might take much longer than they think.
Adam: I know exactly the situation you mean–they’re asking for the moon on a plate, but they don’t realize it. The classic example is someone coming into the forums after they try out Mandriva, and it’s the first Linux they’ve tried.
They say “well, why doesn’t my wireless card work?” and “why doesn’t this USB Bluetooth adapter work?” and “Linux is just rubbish! Why doesn’t all this stuff just work straight away?”
I find all that a bit funny, because I came in around 2001, when you’d have the exact same tenor of post on mailing lists or Usenet at the time. The difference is that back then, it would be, “Why doesn’t my network card work? Why doesn’t my graphics card work?” We’ve gone from broken graphics cards to broken Bluetooth in five years, which I think is pretty good, actually.
Adam: But the thing is, problems are getting smaller and smaller as we go along. So I sort of view this as an encouraging message of progress more than anything else. They complain about things that don’t work and as you say, it’s not their fault, because you have to have quite a bit of technical knowledge to understand why it is hard to implement certain things.
Scott Swigart: You mentioned that you worked on the mobile stuff because it was interesting to you, but you also suggested that isn’t necessarily the way a lot of the Mandriva developers work. They might not have experienced the same pain as some of the other users. What do you think are some of the typical places where the itch that a user has is different than the itch a distro developer has?
Adam: One straightforward example is anything you can do from a console. I work remotely, and I’m more graphical than most of the technical guys on the distribution because I’m sort of moonlighting–like, I work in gedit. So when I go to the Paris office, I see a developer’s desk and it’s ten xterm instances, and you can’t see anything graphical underneath that.
I have another good personal example, even though it’s been superseded a bit because Kate and GNOME can both do it. But back around two or three distributions ago, I noticed that there wasn’t any easy way to launch a graphical text editor as root.
You could just open a file and edit it. You know developers are not going to notice this, because they just do things from a console. It’s natural for them just to go “su vi” or something.
So I came up with just a little menu entry that used consolehelper to launch gedit as root. A lot of people found this really, really useful. I do various things writing documentation on the Wiki. I had to write down “open a console, type su, enter your root password, and then type kwrite or gedit filename,” and I thought that was a bit silly.
Why can’t I just click on an icon and open a root text editor? A developer wouldn’t even notice a little thing like that or consider it necessary, but to users, it’s quite a big deal.
Scott: I got my start on Unix systems, so I’m used to the ten xterm windows, and I remember kind of feeling like GUIs were garbage to some degree. The command line was fast, and you could do exactly what you wanted.
You could pipe stuff to grep, you could filter it, and it was just way faster than clicking a million text boxes to say what you wanted. But that’s fundamentally different from what users have been conditioned to expect with Macs and Windows and that kind of stuff.
Talk a bit about the evolution of developers coming to terms with understanding the idea that users just aren’t going to be interested in learning a lot of command lines to be productive and happy on Linux.
Adam: That’s an interesting area, and it’s something that Mandriva has been on top of for a long time. If you go back a ways to the Mandrake days around 2000 to 2003, we were the distribution that did this kind of thing.
Mandrake was the big one, because we had the Mandrake Control Center, the Mandriva Control Center, which is a big graphical thing that lets you configure a lot of the system. That has always been one of Mandriva’s strong points, but it’s definitely evolved over time.
It used to be that a lot of new Linux users who were not top-level technical gurus wanted to open up a console and learn how to pipe stuff to grep and so on.
Now, a lot more of the new users are people who genuinely don’t want to learn that stuff, and it’s possible to run a modern distribution–not just Mandriva but a bunch of others as well–without ever having to do that kind of stuff.
That is definitely a development within the last two or three years. I think it’s fair to give Ubuntu a bit of credit for that, because they put out a graphical product that you install and then feel like you can actually use it as a graphical modern everyday desktop.
Prior to 2005, a lot of people tried Linux for a day or two and quickly realized that they weren’t actually going to be able to use it without learning a whole bunch of technical console crap. They saw that, and they rejected it.
Scott: Right. And Shuttleworth has always been happy to don his flame retardant suit and make himself absolute flame bait. He’ll post something on his blog, and the Linux community will absolutely eviscerate him and say, “I don’t even understand whether this guy gets Linux.” But it gets people talking.
Adam: Yeah, I do think he’s good at that. I’ve never met Mark. I know Jono Bacon at Ubuntu, and he is a great guy. He says that Mark’s a great guy and I’m sure he is, but he does have a great naive act where he says, “Well why can’t we just click a button and do this?”
Everyone’s like, “Well that’s insane. There’s never been a button for that!”
Sean: There’s never been a button in there. There shouldn’t be a button there. And he goes, “Well everybody wants a button to be there.” And everybody eventually goes, “Yeah. OK.”
How would you contrast your business model and approach to packaging and distribution for users and for businesses versus Ubuntu?
It seems that some of what you do is very similar to Ubuntu, in terms of the way you position yourselves, the way the distro is built, and the audiences you target. At the same time, another portion seems more similar to Red Hat, in the sense of your enterprise focus.
Adam: Mandriva has a very wide presence; we’ve always been an “absolutely everything including the kitchen sink” type of distribution. I guess it’s kind of an historical thing, because as I say, we started off a long time ago.
The Linux world was much smaller then, and we were at the time a comparatively large company. It was pretty normal and feasible to take 30 guys and have a web server and a distribution that worked as a server, a desktop, a firewall, and just about anything else.
We are trying to make things a little more focused these days. For the 2007 release, which would be about two years ago, we dropped several versions of the distribution that were editions. We dropped the PowerPack Plus and Discovery commercial editions.
Now we only have one commercial edition, and we sort of simplified the rest of the range. The idea is to focus a little more on Mandriva Linux as mainly targeting a desktop user base and then to separate out corporate server and corporate desktop a little more as a separate corporate line.
The Red Hat comparison probably works best as sort of a mini Red Hat idea. We have the Corporate line and then we have the Mandriva Linux line, which is more an end user community based kind of thing, a little bit like Fedora.
Sean: This was the year of the Linux desktop. Do you feel that one of the things that inhibits that is the fact that every piece of the Linux desktop wants to “market itself” as independent? Users get kind of confused because they get the platform, and Vista hasn’t exactly rocked the world either.
People go in and buy a box that says Vista on it, and it doesn’t say “Volume Shadow Copy version 5.8,” which relieves them of the clutter of feeling like they have to understand whether 5.8 is better than 5.7.
I wonder sometimes whether it’s important for the Linux desktop to fuzz the boundaries of the distribution, sort of the way Ubuntu has done, in order to relieve the user base of that kind of adoption complexity.
Adam: This is definitely my personal view and not the view of Mandriva, but I don’t think there’s going to be a year of the Linux desktop. I think desktop computing is going to become mostly irrelevant before that ever happens. Everything is going to go to mobile devices.
There’s going to be some kind of convergence in the space between netbooks and cell phones and smart phones. We’re going to wind up with a science fiction vision of the device in your pocket that does everything. But that’s just my personal view.
In terms of your question, though, I agree that what you describe is a drawback in a sense for the Linux desktop. I kind of gave up focusing on how to win all the users in the world a couple of years ago.
Adam: Instead, I focus on how to keep making the product better, and how to serve the users we have. I think it’s legitimate to view Linux as something for maybe 20 percent of the most technically inclined users who like to tweak and are curious.
That leads one to the view that maybe what we have isn’t perfect for everyone in the world, but it’s a good product in its own right. It does what our users want, and we’re just going to keep refining it and focusing it and improving it and growing that group of people gradually.
Everyone knows that Apple is never going to have more than 10 percent of the desktop market or something, but no one seems to think that’s a problem. That’s what Apple is supposed to be.
Scott: We could probably talk about Apple all day, especially in terms of the Apple enthusiasts that complain about Microsoft not being transparent.
Adam: Right–Apple is 10 times worse. Apple is the black box, and you have no idea what the hell is going on inside. What comes out is usually pretty good, but if you want transparent development, Apple is not the place to go.
Scott: We had a conversation with somebody a while back about whether proprietary software will continue to exist, or whether it will all go open source. Their point of view was that there’s always going to be room for a great product, regardless of how it’s developed. It sounds to me like the goal with Mandriva is really to be a great product, rather than to espouse open source.
Adam: Actually, we do care a lot about open source, and it’s important to stress that. Otherwise we get a lot of hate mail.
Adam: We’ve always had a free distribution. Since 1998, we’ve been releasing Free with every release. Everything we develop is always free software. We do believe in a lot of the arguments that free software development is inherently a better way to do things. If you go too far the other way, though, you end up with GNewSense, which the FSF loves, but no one uses.
Most people want a great product. If you develop a great product then people are going to use it, and we definitely focus on making Mandriva a great product. We love open source, and we try to push open source development. We got things like RadeonHD and Nouveau in the distribution, trying to push for open source components in place of proprietary ones.
Scott: Open source is a collaborative way of developing. You’re less likely to produce features that nobody’s going to use, because if there really were such things, nobody would write them. Users know that they’re unrestricted in how they use the end product. In other words, basically, my personal view is that the ethos of open source delivers value to the end user, even if they don’t buy into the philosophy of open source.
Adam: I have an interesting take on it because I work in this little nexus. As I said, I interact with users directly every day. I maintain stuff for the distribution, but I don’t code. I’ve never been an engineer, even if I’ve learned how to fix tiny little things just by trial and error and using Google.
But I can actually see everyday occasions where open source helps me and helps the users, even though neither of us have a clue how to write a garbage allocator. There are still instances where it’s important and really useful to me to be able to grab the code or do other things that are directly possible because of the open source nature of the environment.
You can talk to authors, you can download a project, and you can look at the documentation. You can email the author, and he’ll probably get back to you within a day. You’re not going to get anywhere trying to do that at Microsoft–you can try to email the author of Windows Media Player, but it’s just not going to happen.
I see positive outcomes from things like that literally every day in my work, where open source helps people directly.
Scott: I think we’ve come to the end of our time. Thanks for talking with us today.
Adam: Thank you–it’s been a pleasure.