It’s been hard to read any of the open source Planet aggregators or check Twitter recently without catching references to the Linux Hater’s Blog.

I read it on occasion, and generally tend to agree with about 50% to 70% of any given post on the LHB. For example, now that Jeremy has drawn the anonymous blogger’s attention, he decides to rip into his issues with Samba, which mostly boil down to it’s a pita to configure.

Of course, my first response would be that Samba is pretty easy to configure using YaST’s Samba module, but the larger point he makes is valid:

If you want to provide a feature as part of a platform, you can’t just write the code and say “here’s the config file! have at it!”. When you do that, you get a bunch of other people who don’t really understand the configuration space writing config tools for you that all suck in one way or another. If you want to introduce a core feature to the desktop “platform”, you really have to make vertical integration easy, or even do it yourself. Samba, as far as I can tell, hasn’t done this. They’ve written a solid engine with tons of features, but have ignored the the core problem that very few people understand how to configure the damn thing. They’ve effectively left it to others, and the others have failed. Where as Microsoft? yea, their server code might even be inferior, but configuring shares is pretty damn easy in comparison.

Which really brings me to the main topic of this post: the silo effect. As far as I can tell, the silo effect has been around since the dawn of software development. Most software needs to integrate with other software, yet most development teams communicate very little with the teams that write the software they need to integrate with. Perhaps samba could talk to the kernel people, and the fuse people, and the freedesktop people, and the kde people, and the gnome people, and engineer a top to bottom solution that makes samba usable for desktop users. Sure. As some of you flamers like to say, just wait until 2020, and it’ll be there.

As far as I can tell, in the OSS world, the silo effect seems to be magnified. Which is ironic, since one of the things OSS was supposed to do is kinda solve this whole silo thing in the first place. Everyone can see each other’s code. Everyone can modify each other’s code. If the developers of some project didn’t care about your goals, you could always take their code, and do it yourself. Right? There’s open mailing lists and bug trackers so that communication is as easy and as smooth as possible, right?

I do think some of this is going on, but maybe not fast enough. Once you get to the user interface where it’s time to configure a service, things fall down quite a bit. (If this wasn’t true, I wouldn’t be able to list “Apache” on my resume as something I have experience with — who’d care? Obviously, it’s recognized that Apache and the other FOSS server software I’ve worked with professionally is, to be diplomatic, involved enough to manage and configure to count as a skill.)

So, I hope the hater keeps on hating — and I hope that FOSS proponents and developers listen up. It hurts a bit sometimes to hear that your favorite project isn’t seen as perfection itself — but Linux needs some constructive (and LHB is mostly constructive criticism, even if it’s a little sharp around the edges) criticism.

What’s really encouraging to me is the number of Linux and FOSS folks who’ve embraced the LHB instead of trying to deflect criticism. I think that’s healthy and a sign that the overall community is maturing. (I’m not sure the LHB would have been embraced a few years ago…)

I did a quickie interview with Brian over on OStatic last week, but I wanted to also give him a shout-out on this blog. Brian is the newly minted community manager and editor for the LSB Developer Network with the Linux Foundation.You probably know Brian best from his work over on Linux Today.

The LSB Developer Network has a lot of potential to help Linux reach more ISVs, so I look forward to seeing Brian whip it into shape.

Mozilla’s Christopher Blizzard has whipped up something pretty cool — Whoisi, a search engine/tracker to follow your friends’ feeds and other online output.

So, for instance, if you want to follow Blizzard’s online trail, just go to his Whoisi page and you can see his most recent updates on Flickr, Twitter, Identi.ca, LinkedIn, etc. (Or you could check out my page, or just hit up a random user.)

I’d encourage everybody in the openSUSE community to set up a page for themselves if they have a blog or other online presence that Whoisi can track.

This week I’m working off an EVDO connection while I wait for Verizon to get FIOS set up in the townhome I’m moving to. (Just moving across town.) Now, a few weeks ago I blogged about setting up EVDO using wvdial on openSUSE 10.3, so I thought I would mention the process for 11.0.

I hadn’t used the EVDO card in a few weeks, so I plugged it in on my big green laptop that I’m running openSUSE 11.0 RC 1 on. Under GNOME, when I plugged in my EVDO card and went to NetworkManager, it had already recognized the card and all I had to do is click “connect.” No need to give any information (I guess the card stores it all?) — just plug it in and go.

This isn’t just “user friendly,” this is significantly easier than using the same card under Mac OS X or Windows. No (extra) software to install (the management software is a PITA on OS X), it just dials up and goes. Beautiful!

It’s unbelievable how far Linux has come since I went to using a Linux desktop full-time in 1999. From “stable but hard to use,” to “stable and easier than other OSes,” in just nine years.

Update: Should point out that NetworkManager works the same way for users with UMTS, so the network goodness is not limited to EVDO. Anyone have a UMTS card that they’d like to lend me next time I’m in Europe? I think I should test this extensively… (and having a convenient network connection while I travel would just be an added bonus…)

Just did one of the first (I hope of many!) interviews for openSUSE 11.0 coverage. Even though I’ve been with Novell since February, it’s still a bit odd to be on the other side of the interview. I’m so used to asking the questions…

It was a good interview, though — it’s obvious the reporter had done his homework, none of the questions were total stumpers (”So, can you tell me how many kernel header files have changed since 10.3?” “Uhhh….”) and I think the story will be a good one.

A couple of things I mentioned:

• What’s new in 11.0 from 10.3
• Improvements to the installer and packaging system
• Upcoming openSUSE Build Service 1.0 release and how that will affect openSUSE
• How we’ll measure success (metrics we use to see how openSUSE is growing)
• Hurdles to success (convincing new users to use openSUSE / something other than Windows)
• Roadmap for 11.1, etc.

So, my question to the audience — if you had a chance to talk to a reporter about openSUSE 11.0 (or just openSUSE in general), what would you want to say?

What Linux desktop is most popular?  Ubuntu, openSUSE, Fedora, Mandriva, Slackware, or another distro?

Steven J. Vaughan-Nichols writes that he thinks it’s Xandros. Yes, that’s an unexpected result, but he has a point — Xandros is the distro shipped on the Eee PC, and by SJVN’s count, they’ve pumped out about 1.1 million Eee PCs and are still going strong.

Now, by my count, that puts Xandros in about 4th place, but closing rapidly. Ubuntu, unfortunately, doesn’t track — or at least doesn’t release — their installs, but Mark Shuttleworth has been quoted a couple of times as saying that they have 6 million or 8 million users. But they don’t have hard numbers (that I know of) so if we’re conservative and reduce the quoted numbers by half, that still gives us about 3 million Ubuntu users — well ahead of Xandros.

Fedora’s most recent stats say they have about 2 million Fedora 8 users in six months, and our tracking says we have at least 1.5 million openSUSE users on various releases. All of which is a long-winded way of saying that 1.1 million users is impressive, but Xandros probably isn’t in 1st place. Yet.

However — I think the Eee PC’s runaway success shows that the UMPCs are going to be Linux’s path to the mainstream desktop, and we need to be thinking very heavily about that. (Note that we’re on several UMPC devices as well — like HP’s Mini-Note systems that actually have a usable keyboard…)

The new wave of UMPC devices and new mobile devices are a strong opportunity to put Linux in the hands of new users. The resource requirements mean that Linux is well-suited for these devices, and the fact that Linux can be heavily customized also means that Linux is a great choice for manufacturers looking to differentiate themselves from the competition in ways other than hardware and pricing.

But Xandros’ strong showing on UMPCs hasn’t put it in the lead just yet, at least not by my count. Stay tuned, this is going to be a very interesting year for Linux on the desktop…

The FOSSCamp idea is, from one angle, fairly dicey — you take a bunch of people interested in topics that are generally similar (in this case, free and open source software), mix well, and allow them to come up with an agenda that’s their own and which will (one hopes) be productive.

I’ve always been an advocate of a strong “hallway track” for shows — i.e., ample opportunity for attendees to mix it up outside of a structured set of sessions. I’ve been to too many shows where you start to feel like a hamster on a wheel — too busy running from one session to the next to actually stop, breathe, and chat with the other attendees and presenters.

Yes, it’s wonderful to listen to experts talk about the topics they are expert in, but one can only stand so much being talked at without any time to talk with people. At least that’s true for me — maybe I’m just scarred from too many years in college.

But FOSSCamp is essentially all hallway track, right? Maybe that is pushing it a bit too far in the other direction.

In some cases, that might be true. However, it seems that this FOSSCamp managed to get the balance right. I attended several sessions that seemed very useful — including one where the original session didn’t happen due to the session leader not showing up, so we eventually just wound up having a discussion (mostly) about communication between projects, how best to provide software to a community for testing (I admit it, I couldn’t help slipping in several plugs for the openSUSE Build Service here…), and the best ways to credit other projects for work included in a distro (for example, the need for acknowledging where features come from, and how best to give credit).

So, big bonus for the FOSSCamp model of allowing attendees to build their own conference — this works when the attendees are motivated to be involved and share knowledge.

(In the interest of giving credit, I should mention that FOSSCamp was sponsored by Canonical/Ubuntu, prior to their developer summit this week.)

We also had good planned sessions Saturday morning around Packagekit and Saturday afternoon around distro collaboration. There were also good sessions on Friday, but I think Saturday was the better day — after all involved had one day to get acquainted and (for me) a day to overcome jet lag a bit.

In general, I think the FOSSCamp model is a really good one. I’d like to see more conferences have at least one FOSSCamp style day — probably best if served up in the middle of the conference so that attendees have an opportunity to get settled, get acquainted a bit, and confer a bit on what sessions most people might benefit from/like to attend.

If Oracle had its way, there’d be one Linux distro — but who would do the development? According to this post by Paula Rooney, Oracle’s Edward Screven says that Linux distro vendors should compete “purely on the support side of the business.”

That, of course, is complete nonsense.

The tension between Linux vendors to bring in customers through added features and continual development is what helps the Linux community move forward. Without that tension, Linux wouldn’t have matured as quickly as it did, and it wouldn’t continue to improve at such a rapid pace.

I’d also like to know which company is supposed to pick up the burden of development? Oracle isn’t doing it — they’re contributing to Kernel development upstream, but not doing much in the way of advancing Linux beyond that. While the multi-billion dollar company hitches a ride on Red Hat’s development infrastructure, Red Hat, Novell and other Linux vendors are investing in the future as well as supporting the here and now. Very nice for Oracle, not so nice for the rest of the community.

Even though Red Hat might be thrilled at the prospect of being the only Linux, I doubt they’d be thrilled about carrying the sole burden of developing everything to allow companies like Oracle to ride on their coattails.

If Novell adopted the Oracle model, we’d be able to save tons of money on development. We’d also be failing to hold up our responsibility as a Linux vendor to contribute to the foundation of our success. We’d also give Red Hat less incentive to innovate and work on new features. And we’d definitely be less interesting to our customers.

I also don’t fancy the idea of a single vendor in control of the operating system. Even when it’s open source — having multiple distributions is, while admittedly more challenging from the ISV standpoint — better for the market, and better for each vendor because they are not solely responsible for the entire development ecosystem.

Linux needs more contributors, not fewer contributors, and I don’t think all OS development or decision-making should rest in the hands of a single vendor. It’s too important to leave with one vendor or project. This is why I chastized Sun the other day for continuing to tilt at the Solaris windmill.

The Linux vendors do need to find ways to make it easier for ISVs like Oracle to target multiple Linux distros — and that’s why we support the Linux Standard Base — to find a standard that allows companies like Oracle to more easily support multiple distros, without doing away with actual development and advancement that Linux vendors provide.

Sun finally pushed out its Project Indiana yesterday, in the form of a packaged version of OpenSolaris that looks quite a lot like a Linux distro — minus, of course, the kernel that gives Linux its name. On the one hand, I’m pleased to see any FOSS project moving forward. On the other hand, I’m wondering what problems Sun can solve with OpenSolaris that it can’t solve by participating in the Linux community?

The OpenSolaris release was timed to coincide with Sun’s CommunityOne conference yesterday, where I was invited to give a talk on openSUSE and appear on the distro panel with community leaders from Ubuntu (Jono Bacon), Fedora (Karsten Wade), and OpenSolaris (Glynn Foster) and moderated by Sun’s Barton George. (We also recorded a podcast shortly thereafter, which should be released in the next few weeks.)

Removing redundancy

The panel discussion was probably one of the most lively, and certainly the most fun, that I’ve been on. One of the things that we all seemed to agree on — there wasn’t a lot of strong disagreement on any topic, believe it or not — is that there’s a great deal of room for additional collaboration between projects to remove redundancy between distros and the efforts to fix bugs in upstream projects.

Glynn noted a few times that, when wearing his GNOME hat, he sees a lot of waste in different Linux distros fixing or developing things internally with respect to GNOME rather than doing so within GNOME itself.

I think we’re (we in this case being the Linux community in general) getting better in this regard and I’ve observed a lot more willingness to collaborate between distros and projects like GNOME and KDE in the past year than ever before. As Karsten points out “It turns out that, surprise surprise, for the vast majority of concerns, we are all in violent agreement. When you get down to where the differences exist, the less than 5% area, it is hugely gray.”

However, Karsten also compares the panelists to football captains “and it turns out that, aside from the uniforms and different team names, we’re all footballers and think the game should be played the same way.”

I think that’s true, with the exception of Sun — we’re all suited up to play “Linux,” and Sun’s taking the ball to its own field. While Sun collaborates in many of the same communities that Linux vendors and projects do, I’m a bit disappointed that Sun is trying to replicate the Linux experience with OpenSolaris while remaining separate and trying to build a parallel community to ours.

To abuse Karsten’s football analogy further, it’s like the difference between what the United States refers to as football, and what the rest of the world refers to as football. Similar in the respects that each game has a community, but the players are not interchangeable.

Control

Ultimately, I can’t help but think that the problem that Sun is trying to solve with OpenSolaris is not a technical one, but one of control. Specifically, the company is not ready to cede control over its operating system to the community at large, and so it is instead trying to build a community around OpenSolaris rather than joining the larger Linux effort.

If you look at the design decisions behind OpenSolaris — a GNOME-based desktop, an APT-like package manager, timed release cycle — Sun is clearly trying to ape Linux without actually being Linux. Perhaps Sun thinks its operating system is technically superior, and I’m sure there’s some of that — but the company could choose to license its software in a compatible manner and attempt to bring its benefits to Linux rather than pursing a silo-ed strategy.

Before the objection is raised, I’ll concede that control is also an issue for Linux vendors, but to a much lesser degree. As Sun’s Ian Murdock discovered with the failed DCC Alliance, major vendors and projects weren’t quite ready to cede control over the base distribution — but we all have the same general goals, and the work that Novell, Red Hat, Canonical, and others do towards Linux usually benefits the entire community. I don’t see a great deal of benefit in trying to create a parallel community controlled by a single vendor.

Even if Sun is successful in building a strong external community around OpenSolaris, it will take years of effort, and I think that Sun would benefit much more by concentrating on Linux than by trying to clone the Linux community with itself at the center.

I think that it’s in some sense ironic that Sun made its OpenSolaris announcement at a conference titled “CommunityOne” — community is in one sense about the individual members giving up some control in the interest of the greater good. It’s something Linux vendors are starting to learn, finally, and I hope that one day Sun learns it as well.

Novell is working with Microsoft to help Microsoft cooperate and contribute to open source projects under open source licenses, and sticking with the existing standards to do so. Keen observers of the tech industry might note that this is a bit of a rarity, but it’s something that I hope we’ll see more of from Microsoft.

The announcement went out today that Novell and Microsoft are collaborating around the OpenPegasus project and other system management tools.

Thanks to Novell, Microsoft is going to be contributing to several open source projects — and making Linux easier to manage. Yes, you read that right. It will also make Windows easier to manage using Linux tools — which is going to be a breath of fresh air for Linux admins tasked with managing Windows boxen as well.

I spoke to Eric Anderson, vice president of engineering for Novell, about the agreement and got some additional details on how the deal came about, and a little background. Eric says that Novell was approached by Microsoft for help working on these projects and that we happily agreed to do so, as long as Microsoft work with those communities and contribute back to them.

Of course, one of my first questions was “what licenses will Microsoft be contributing under?” The OpenPegasus project itself is licensed under the MIT license, and other parts of the project involve providers under the SBLIM which is under the Common Public License, and providers that are under a BSD license.

Microsoft will also be submitting some code under one of its OSI-approved licenses.

Why this is important

Getting Microsoft involved, as an equal participant in the project, is a huge step. Microsoft has started to take tentative steps into the open source community, but this is one of the first times that Microsoft is going to be participating on equal footing with the rest of the open source community on a project started externally.

I may be an idealist — OK, I am an idealist — but I think nothing but good can come from getting developers to sit down together and work on common code — even when, maybe especially when, one of the parties at the table is Microsoft.

It’s hard to maintain a hard-line stance when you sit down and work with another person on a common problem. I don’t expect an overnight cultural change at Microsoft as part of this collaboration, but I think it will have a positive effect. The only way to find out is to try.

This is, of course, good for Linux and the community too — anything that makes it easier to manage Linux boxes along side Windows (or in any heterogeneous environment, for that matter) is going to help Linux adoption. Giving Windows admins tools to easily administer Linux boxes will help to remove some of the barriers to adoption in Microsoft environments.

I talk to admins all the time who express frustration in getting Linux into their environment, who have to fight arguments against Linux because it lacks the standard management tools that come with Windows. Well, that argument will be easily dismissed in the not-too-distant future, and I think that will help accelerate the adoption of Linux in the enterprise even more.

Of course, this is good for Microsoft as well. Its customers have been asking for Linux, and by providing better Linux support, Microsoft makes its customers happy, and helps Microsoft remain relevant with those customers. But this also means that Linux will continue to spread through the enterprise, and that’s not a bad thing at all.

It’s not just Novell and Microsoft in the management boat, though — this is something that has been a sore point for Linux for years, and we’re hoping that the community can work together on this front to make Linux management easier.

Eric stressed the opportunity in this arrangement for the Linux community to help improve Linux management capabilities as well. Eric says he’d like to see the Pegasus community work on writing management providers so that every service (where it makes sense) is manageable through this framework and so that the Linux management experience is as cohesive, powerful, and easy as any other platform on the market.

I couldn’t agree more — the easier it is to manage and work with Linux systems, all the better for Linux adoption.