Nat Friedman: "Flamewars are part of the community culture"


    Novells CTO / Open Source doesn't expect long term problems from the Microsoft deal - an interview about the future of the Linux desktop, legal DVD-support and the mistakes of the Hula project

    Nat Friedman has been one of the driving forces behind the development of the Linux desktop for a few year now. First with his own company Ximian, founded together with Mono chief architect Miguel de Icaza, after its acquisition now inside Novell. A few months ago he has been named "Technologist of the Year" by the VarBusiness magazine for his work around the SUSE Linux Enterprise Desktop. Since then he has been promoted to "Chief Technology and Strategy Officer for Open Source", besides the desktop he is also overseeing Novells server business now.

    During Novells Brainshare Andreas Proschofsky had the possibility to sit down with Friedman and talk about the Linux desktop, the consequences of the Microsoft agreement and the mistakes of the Hula project.

    This interview is also available in a german translation. What are the most important improvements you are going to introduce with Service Pack 1 for SUSE Linux Enterprise Desktop (SLED)?

    Nat Friedman: Service Pack 1 really is no big feature release for us. One of the interesting things though, is the work we have done closely with customers - like the Peugeot deal - where they actually deployed SLED in their production environments, we have learned a lot from that. Resulting from that there has been lot's of interoperability work, Active Directory, Printers, cause we've seen that this was a major issue. We also did a lot of work on Microsoft Exchange support, which is significantly more mature than before now. Did you consider to do a MAPI-connector (like Outlook itself does) for Evolution instead of accessing Outlook Web Access?

    Nat Friedman: We did, but MAPI is really huge and complicated, and not everything is documented properly, so this would be a lot of work. The way we do it now is using Webdav, which is much easier, cause the interface is documented. it is slower, but we think it works well enough for what people want to do. Seeing that the Groupware-Client Evolution is a major piece of the desktop, it doesn't look like Novell has too many developers left working on it, implementing new stuff.

    Nat Friedman: That's not true, Evolution is handled by our Bangalore team. there is nothing at all that changed in that respect for the last 18-24 months now.. But there is a certain lack of external contributors...

    Nat Friedman: No, I wouldn't agree to that either. There's a couple of people working in the community and also in other companies, for instance Scalix. Sure, they are mainly doing there own backend, but while they are doing this, they also end up doing work in the core.

    And there are also the people in the community who are repurposing the Evolution code for their own stuff, like the Contacts and the Dates applications which make use of the Evolution Data Server (EDS), which is great, as such things are the reason, that we split out the EDS from the Evolution Core in the first place. They might even be drafting the future user interface for Evolution by doing that, cause they are free to experiment, and if that's interesting we might eventually replace the Evolution UI with one of those.

    Another example is the Tinymail-work that Philip van Hoof has been doing, as a result of which he has been doing huge performance improvements to our mail access library Camel.

    Picture: Contacts, one of the applications utilizing Evolution Data Server But most of that optimization stuff hasn't gotten upstream until now.

    Nat Friedman: Well, there is a lot that hasn't cause the guy is incredibly productive, so it's difficult to take that all, but there is also a lot that has.

    So all in all we've a huge 300.000-400.000 line application, we have a dozen or so engineers working on it inside Novell, and a bunch of others in the community and in other companies, so I feel pretty good about the work that is going into it. Coming back to Service Pack 1...

    Nat Friedman: Besides what I've already talked about before, there are also some usability changes, like changes to the main menu, a new world clock and so on. Most of the work has been hardware enablement though, so there is tons of new printers supported now for instance. Also Lenovo laptop supported improved dramatically, as Peugeot is mostly using Lenovo X60 and T60 laptops, so for instance hot docking and undocking works now.

    Picture: SLED SP1 features an update main menu Why did you go for a new rewrite of the main menu?

    Nat Friedman: It's not a complete rewrite. But there are two reasons for the changes we've made: One - we got a lot of user feedback, which got used to it and made some suggestions for improving it. We also realized that some stuff didn't work as planned, like switching to favorite applications, so we made this much more discoverable by putting the two most recently used applications just at the bottom of the main view.

    Also the stability is a lot better now and the memory usage a lot less. But the application browser still takes some time to open up in the first run, which seems a thing a lot of people are complaining about as this basically breaks your workflow.

    Nat Friedman: You know, Service Pack 1 is still not out, so there is still some work left to do. We have set some performance goals for those kind of issues, both the main menu and the "more applications" should open in less than a tenth of a second.

    Besides that, we've also updated some of our core applications, like going with Firefox 2, Flash 9, an upgrade for the Acrobat Reader, obviously lots and lots of changes in Compiz. The reason I popped up a list of 350 items during my keynote is because that's what SP1 is. It's not one or two big things, but a lot of small improvements.

    Picture: The new world clock for SLED SP1 But still, that's a lot more than you normally do for a service pack?

    Nat Friedman: Yeah, that's true. To us, we consider SLED 10 sort of a first generation product, so the Service Pack is a chance to solidify that and fix some things we learned from our customers. With the original release of SLED 10 you introduced the Real/HelixCode backend for you music player Banshee, which allowed legal MP3-support to be shipped with the product. As far as I remember, there has been some discussion about also doing this for some video codecs, for instance for the Windows Media-stuff. Is this still being worked on?

    Nat Friedman: Yes it is. We wanted to get this for SP1. Basically we are waiting for Real Networks to deliver working Windows Media Audio and Video decoding to us and they are behind schedule. So as soon as they deliver that to us, we'll be shipping that to our customers, independently of SP1 release.

    Picture: The music playing application Banshee now has podcast-support but is still waiting for WMA-codecs from Real Did you also think about supporting DVD?

    Nat Friedman: Yes we did, but the thing is, you need to license MPEG-2 for the video, you need MP3 and Dolby Decoding for the audio, you need CSS, so there are quite a few components. And the most expensive one is actually MPEG-2, which alone would cost something like 2 Dollars per license, given our low pricing, we would have to make this available as a separate buy, otherwise we couldn't make interesting discounts for our customers anymore. After what seems like a quite successful introduction of SLED to the marketplace, did you get some new developers for the Desktop team?

    Nat Friedman: We already have a pretty big Desktop team now, it's over 80 engineers now, very distributed over the world. People working on,, Evolution, Firefox, GNOME and other stuff. But we are also actively hiring at the moment. Given the strong negative reactions lot's of people in the community showed to the Microsoft deal, do you think it'll get more difficult for Novell to get their stuff upstream?

    Nat Friedman: I haven't seen that until now, I mean there have always been flamewars in the Linux community, it's part of the community culture. Sure there are consequences for Novell in the community resulting from the deal, we have seen that, but not in the respect that someone says "Well, Novell as a business did this agreement with Microsoft, so we won't accept their patches". And most of that negative sentiments don't seem to come from the people who accept patches anyway, they come from people who have a sort of "professional commentator" role in the community. But for instance there seem to have been some quite negative reactions to the new main menu (which was proposed for integration in GNOME 2.18 and declined)

    Nat Friedman: Yeah, but let's be totally honest here. I've read this thread and as far as I could see, this was all based on feedback, like "It uses to much memory" or that it is not mature enough at the moment - a sentiment which we obviously don't share - this has nothing to do with any business decision that Novell might or might not have made. Also there is Beagle, which is more or less reimplemented in C by the Tracker project, just because it was written with Mono.

    Nat Friedman: Sure, but that's the way it always was like under Linux, look at how many X-Servers, how many window managers and so on we have?. That's just a standard thing in the community.

    Beagle has a lot of strengths, it indexes nearly everything, it has extremely fast query speed, it has multiple user interface, a very extensible architecture, there's just so much it does. One weakness historically has been memory use, but this has improved dramatically recently. If you check recent versions of Beagle you are on 25 Megs most of the time, and the developers are constantly improving on that, the goal being to get it around 8 Megs.

    Tracker on the contrary doesn't work half as good as Beagle, it doesn't do half the things that Beagle does at the moment. But that's ok, if people want to implement other the indexers, they can do that, I just don't think there is a need to.

    I think what you'll see as the Beagle memory footprint gets lower and lower, it'll become even more dominant in terms of who uses it.

    Picture: Desktop search application, Beagle Do you have someone working on tagging for the desktop?

    Nat Friedman: Joe Shaw of Beagle-fame is working on that, he has recently blogged about doing a Metadata-store interface for Beagle, which is exactly what this is, allowing tagging. That was the one feature that Tracker had, that Beagle hadn't and now Beagle has that as well. Are you interested in doing something like Time Machine / Shadow Copy, were older versions of you data can easily be restored.

    Nat Friedman: Yeah, I think that's interesting, it could be very good to have something like that, I'm not sure exactly how to do that. That's something we might think about for SLED 11. Are you going to look into a deeper integration of the desktop and Web-Services

    Nat Friedman: There are a couple of areas where people are doing research on that, and I think we are going to track that research and pick up the right parts. So for example one of the next steps in the evolution of Gimmie is integration with Web-Services. So for instance you can see if someone is online via their Flickr-Account.

    Also I guess the Mugshot stuff that Red Hat is doing will go into this direction, I mean they have been doing all this Web-Services stuff, they got desktop expertise there, so it would be just a natural next step for them to start kind of Mugshot-like-services in the desktop.

    Picture: Mugshot What ever went wrong with Hula?

    Nat Friedman: Well, we learned a couple of lessons there. The biggest mistake we made is that we didn't make a release, which made it difficult to build a community. I guess the sense we had, was that what we are doing is so great, that by the time we release it, everyone will be so blown away by it.

    A good thing that has happened now, is that one of the Novell customers, Messaging Architects, purchased Hula and is continuing to run that. I'm actually quite excited about that, cause there is some really fabulous technology in Hula, like the AJAX-interface.

    Also there is a fork called Bongo now, were they are doing some great work, so I guess Hula will live on through both of that. Thanks for taking your time to doing this interview.

    (Andreas Proschofsky)



    • Bild nicht mehr verfügbar

      Novells "Chief Technology and Strategy Officer for Open Source", Nat Friedman

    Share if you care.