22 February, 2011

Priorities: communicate and be a mentor!

I just reviewed and published the news.o.o article on openSUSE's GSOC participation. We're seeing quite a few ideas being added to our Ideas page on the wiki and some mentors have stepped up as well - but we need more. Seriously, if students are to pick openSUSE as an interesting organization to do a GSOC project, they need to have some exciting ideas! And mentors who can help them get the most out of their 8 weeks of hacking.

for openSUSE

For openSUSE, GSOC is a great opportunity. I've already heard marketing people talk about how cool it would be to have an Android app for people to read the Weekly News or follow planet openSUSE easily. Or how cool it would be if SaX, the well-known and missed-by-many X configuration tool for openSUSE could be resurrected (there is an idea and a mentor has stepped up already!). The build service team has added some ideas too but I bet other areas of openSUSE could also use some help. Having a student work 8 weeks full time on something which needs to be done is nothing to be sneezed at - even though it takes some time for a mentor to help them out.

But frankly, in my opinion the value of the code itself is nothing compared to the opportunity to gain a new contributor for openSUSE. Many students stay around in the projects they wrote code for - Cheese wouldn't exist (and be maintained!) if it wasn't for GSOC. And we all know how popular that app is at demo booths around the world ;-)


But we have the openSUSE 11.4 release come up! And once that is over, there is bugfixing to be done, work on the next release... We're all busy and not all of us might feel capable of helping a student out, or want to spend time on doing that. Isn't it more important to make sure 11.4 is a rock solid release?

I say no. Seriously. As any Free Software project, openSUSE gains and looses contributors all the time - people join but also leave because they get jobs, become to busy, loose interest or otherwise aren't able to contribute anymore. It is vital for us to keep working on getting new people involved. So if you have a choice between fixing that bug in Plasma NetworkManager for the upcoming release, upgrading that GNOME package in time, fixing OBS - OR spending time on mentoring.

What do you do? Think about the long term. Surely, having bugs in our release is bad. Most of us surely qualify as perfectionists and see the quality of their work reflect upon them. And having show stoppers chases users away. But. I am sure you have to agree with me that openSUSE will never be perfect. There is always another bug to fix. Meanwhile, in another 8 months we'll again be close to a release - and you'll have to do your work alone if you don't get help! It is surely more fun (and more effective) to be squashing bugs together than alone and to make that happen, you have to do something.

Blood, sweet and tears

You all put in your hard work in openSUSE and you all want it to succeed. So, you do that by working harder? Maybe. I don't deny that hard work is where it starts. But it is NOT enough, believe me. The difference between successful projects and unsuccessful projects lies rarely if ever in technological superiority or even in the hard work of the developer who started it. You can't do it all alone, and not until eternity either. For the continuity of your work as well as the quality it is crucial to have others who care about it.

So you have to spend time on helping others and getting them involved. You have to spend time communicating about your project. If others don't know about it, they won't use it nor help out. That is why writing a blog regularly matters - surely, I know many of you don't even have time to read blogs, let alone write them. But like other Free Software projects, we have many people who are interested in what is going on in openSUSE. Blogs, articles on news.o.o and even just mails to -project help them stay up to date and might convince them some day to help out. And once they make that step of coming forward to help, they need to be mentored. Be it for a GSOC project or in their own free time, it will be a challenge for them to really get into it.

We don't realize how difficult it can be to get involved. How big the step can be to set up a development environment, let alone send in a patch. Or even just ask for help. So you have to make that easier. You have to offer help. Write documentation, make people enthusiastic about what you do by communicating about it. And by making time for them when they need it. Even if there are urgent bugs - people matter more than code.


You might think you can't do it. Or that it won't help much. Maybe you've been disappointed in the past by someone who you've tried to help and it didn't work out. Well, get over it. Life sucks and things don't always work out. But if you give up, you don't just give up on some potential help. You give up on your project.

Sometimes you can blog many times asking for help and nothing happens. But I have also seen a simple call for help ("we need a developer to help out on our website, just mail me") result in 5-10 people step up. I am not kidding, almost 10 contributors in a week time after a single blog. And mind you, that blog was not special. It didn't have pretty pictures or a well written text. It was just a cry for help and it worked. And once those people stepped up, answering their questions and working with them resulted in a great product. Most of them are still active!

Step up

So we need you to step up. Decide for yourself: do you care about what you do? Do you want it to survive you and the company you work for? Then drop that bug you are working on and put your name on the Ideas page (HUGE credit to Manu Gupta for all the work he did for planning GSOC for openSUSE!!!). Blog more often, send a status report to -project (and CC the openSUSE news team in case they want to give it even more exposure). Communicate. And make time for anyone who even remotely starts to think about possibly helping you out. Document your code, keep it clean. And make sure it is crystal clear how someone can get involved.

Don't bother with pretty pictures if that is not your thing. Your blog doesn't have to read like an essay. Simple is enough. Perfect is the enemy of good, remember? But step it up. Without communication and mentoring, openSUSE has little future!

18 February, 2011

On getting and giving attention

Dear openSUSE community,

There is a lot going on in openSUSE. Some of those things are reasonably well known by now - Tumbleweed for example. Other things like the cross-desktop MIME handling Stanislav Brabec has been working on are not that public. But should be! Now that is where everyone can make a difference. If you're a developer but also if you're just interested in what is going on! Development of important things happens because people know about it and care about it. They know and care about it because there is communication done. The openSUSE marketing team and in particular those writing for news.opensuse.org see it as their task to help spread the word about things in openSUSE which are important and can use help. But our marketeers can't be everywhere - they don't always know what is going on. So they need help with that.
Speaker Cone


Simple. If you are a developer working on something cool and new, let the openSUSE marketing team know. That doesn't mean you have to write a story - just ping one or more members of the team or mail the mailinglist or news@opensuse.org. They'll get back to you, ask you questions, write about your initiative. And if you're aware of something cool a developer has done or is doing in and for openSUSE but you haven't seen any writing about it, think about letting the marketing and news team know! Send links to mails or blogs where they can find some information or help them contact the developer(s). Obviously, if you can write a quick summary of what is going on and why it is cool - that is awesome. You don't have to write the article, the marketeers can do that for you, but any material you can give or point them to will help them do it!

This way we can more easily give some of the cool new initiatives in openSUSE more attention. openSUSE is a bottom-up community, the initiatives by individual contributors make the difference - that also means we need to work a bit harder to get the word out on what is going on. Help us with that, please!

Linux Starter

On an slightly related note (getting the word out), the Dutch "Linux Starter" Magazine still is and will continue to be available from this site. Yes, you can finally order them now! Go ahead if you want to start with Linux and/or openSUSE, for about 10 bucks you have an excellent source of information and a nice custom dutch openSUSE DVD! Spread the word on that too :D

17 February, 2011

Banshee, referral money and how to earn a honest living

I'm catching up to some reading and bumped into Joe's interesting article on Open Source Report about the Banshee Amazon store in Ubuntu 11.04.

How not to make money?

If you haven't read the article, you should. Its about a question which has been on my mind too. How can a Linux Distribution make money in a sustainable way? This move from Ubuntu takes away resources from downstream - the party actually writing the code they ship. This community, partially funded by Novell, donates the money to the GNOME Foundation. I can't politely express how low it is to take that away so I won't try. However, it does beg the question how one should do better.

Needed for...

openSUSE is in the process of setting up a Foundation (or e.V. or...). Once there is a Foundation, it will be on the lookout for funding. Obviously my employer will support it, we are a stakeholder in the future of openSUSE. And provided we support the Foundation's goals. But the Foundation will also want to explore other ways of generating income.

What doesn't work?

Donations and merchandising don't seem hugely profitable in other communities. Sure, openSUSE did well at FOSDEM, selling 16 crates of openSUSE beer and donating the money to FOSDEM. But we're still talking about a few hundred euro's and that's including the t-shirts we also sold for FOSDEM. That wouldn't keep the openSUSE Foundation running. So I understand that the $10.000 that Banshee brings the GNOME Foundation each year is interesting from a distro point of view. But ethics and common sense should play a role here too. I guess it might make sense to take a 20 or 30% cut in discussion with the projects - not 75 or 100%. So it might bring some revenue. Not enough still.

I've been thinking about the individual sponsorship program - communities like openSUSE, KDE and GNOME have lots of people who used to be active members but since then moved on. The skills they learned while active surely still help them every day. Maybe they even got a job because of their activities! So why not set up an alumni program, target them to give a small share of their income to the community? Organize something for them - a yearly reunion, a special alumni meeting at a yearly conference, some way to share what they have learned since leaving openSUSE...

So what DOES work?

Good question. Maybe the answer is all of the above. Use referrals as much as is reasonably possible. Take a 20% cut of the Firefox/Google referral, 20% of Banshee/Amazon etc. Sell some merchandising. Set up a Individual Sponsor program, an alumni program, try and offer companies incentives to sponsor as well. And do specific, targeted fundraisers for specific causes sometimes. But stay reasonable. An e.V. or Foundation should protect, support, mediate, communicate. It doesn't need to employ 100s of developers or marketing people or consultants. That is for the other parties in the ecosystem.

Ideas are more than welcome

This is something all Free Software projects struggle with. Surely there are plenty of articles about how companies can develop a sustainable business around Free Software _ found less on how communities should generate income. So ideas and links are welcome!

*photos by kaero on flickr*

16 February, 2011

FOSDEM 2011: building distro bridges

FOSDEM. I finally got to the "blog about it" todo I took from there. I have to talk about the distribution collaboration panel discussion Jared Smith (Fedora Project Lead), Stefano Zacchirol (Debian Project Lead) and myself led on Sunday (video here). We discussed what barriers there are to cross-distro collaboration and what to do against them. I can summarize the whole thing by saying pretty much everyone in the room agreed more collaboration would be good and won't be that hard. More specifically, the main arguments and their rebuttals that I heard:
  • We will lose our identity and become one big grey blob!
    Now first of all, grey blobs can be cool. Second, our identity does not lie in any single tool or application - our identity lies in the people, philosophy, culture AND technology combined. Collaboration on a software installer or init system won't change that!
  • We will never all agree on anything!
    And why is that a problem? Maybe openSUSE and Fedora work together on the init system without input from Ubuntu while Mandriva and Arch push forward a new Software Installer without Debian's collaboration. Boehoe. I'd say such a situation would already be a big improvement over each doing its own thing!
  • We all hate each other!
    No, we don't. As was said in the discussion panel, usually our users get a whole lot more passionate than our contributors. openSUSE and Fedora, Debian and Mandriva - we all DO work together already. Fedora and openSUSE collaborate on bringing LXDE patches upstream, KDE and GNOME packagers have worked together since ages. There is plenty good going on already.

It is not that hard

I think the last point makes clear what Vincents presentation on Appstream and the success of the cross-distro collaboration meeting on application installation also pointed out: we're doing it, NOW. The reason the Fedora and Debian PL's as well as myself wanted this meeting and are talking and writing about this has little to do with resistance to cross-distribution work in our communities. There's little of that. As I expressed at the panel, the issue is that when we encounter an issue and start thinking of possible solutions, working with other distributions isn't a high priority. Rather, we just don't think about it. That means we start to build a solution, only to discover later that there already are others under development. And I bet the result is often that all solutions have their pros and cons and none is as great as a unified solution might have been. Too bad! Because if you DO try to collaborate, you might find yourself in the same situation Vincent found himself in: he cautiously proposed "maybe we should try and work something out together?" and it turned out everyone loved the idea!

So our issue is awareness. I think we already did a lot to advance that at our last openSUSE conference: Collaboration Across Borders. It really made a difference and it is where Appstream was born.

A good example

To show how easy it is to communicate, let me point you to this recent email on the distributions mailing list on Freedesktop.org. In that email, Czech SUSE hacker Stanislav Brabec just tells people about a solution for a problem he's been working on: the maintenance of MIME defaults in glib and desktop-file-utils and making it all desktop-specific. This work can help applications integrate better within different desktops. For example, Nautilus in GNOME will start file-roller if the user opens a ZIP archive but if Nautilus runs in a Plasma desktop it will start Ark. If the other distribution people like this work and want to help out, that would be awesome. We would get better integration between Plasma, GNOME, XFCE, LXDE and the other desktop technologies on Linux. And the distributions would share the workload!

Help out

So, things are happening. But you can count on me to keep talking about this and trying to help make it happen wherever I can. Jared and Stefano promised to do the same and I hope other distribution leaders will follow suit. And I count on all of you to keep collaboration in mind! More importantly, do it, help push it, help make a difference. As I wrote before in my way or the highway, collaboration strengthens us all and is what makes open source work so well...

15 February, 2011

Nokia does the right thing!

For those who've had it with the Nokia discussions - further down I have some positive news too...

So the rumors turned out to be true. Stephan Elop, the new Nokia CEO (and ex Microsoft employee), has decided the future of Nokia is to work with Microsoft. It will be possible to decrease investments in R&D and Nokia won't have to build it's own platform nor focus on services anymore. Instead it can focus on becoming a phone hardware manufacturer, a strategy which will ensure high margins!

Sarcasm aside, it seems to me that Nokia (or at least its CEO) has realized it has had its best time in the nineties and it is over. Better to show good profitability for the next 4-6 years than invest in the future. Within 2-3 years Stephan will leave a company which is "going great" (short-term financially speaking) and receive a big bonus. Usual business of the type which brought us the credit crisis.

How about us?

So what does this mean for Free Software? Hard to say yet but I think it's a fair bet that development on MeeGo won't increase due to these choices. Bad for the linux kernel, infrastructure like Telepathy, Gstreamer and top-level stuff like Qt. Qt will probably not be hurt that much, it is the platform of choice for Symbian for at least the next few years. And if they change their mind Nokia (or otherwise some volunteers, can't stop them can you?) will port Qt to WP7 - write one app, run on MeeGo, Symbian AND WP7. With the open governance model of Qt & the LGPL license, the decreased investments could be compensated by community efforts. Still, I feel sorry for those who might get fired over this. And personally, I was looking for a really open mobile platform - my N900 is pretty awesome and though I'll buy the N9 it's sad it won't have a decent successor.

Meanwhile, stock has gone down some 15% on Nokia, MS' stock went down, Apple and Google went up. Hmmmm, what does the stock market think... Anyhow. I guess more than enough has been written and said about it, let's just wait and see what happens next...

Good News - twice!

In far more positive news, my own employer (that would be Novell) has helped the stability of the UK financial system by migrating the London Stock Exchange from a Microsoft .NET based solution to a SUSE Linux Enterprise stack. The new trading system went live last week. The earlier Windows and .NET based software actually managed to get the CEO of the LSE fired after an 8 hour outage (any idea how much money that costs?). Their new CEO, slightly smarter than Elop if I might say so, decided to go for Linux. This has led to better performance and stability - no surprise. SUSE proves to be about 15 times faster than the MS/.Net system which couldn't get network round trips below 2 milliseconds despite big investments in code and hardware improvements. Meanwhile SUSE comes in at about 126 microseconds! I don't expect any 8-hour outages or fired CEO's, so congrats to everyone involved!

And while at it, let me also mention an interview with my colleague Kerry Kim. Subject was IBM's supercomputer 'Watson' which took on a human in the TV show 'Jeopardy'. Watson runs SUSE Linux - as do, Kerry mentioned, 6 out of the top-10 supercomputers in the world. Check a video of Watson out here (or embedded above).

03 February, 2011

LCA 2011 event report

Hi all!

I've recovered from LCA 2011 and took the time to write an event report for the marketing team. You can find it here. A quick summary:
  • Event Details
  • It is a well organized and quite professional conference. Biggest sponsors HP, IBM, Intel and Google. Budget about $700.000, entry fee from $60 to $600 (depending if you're student, hobbyist or professional and if you get early bird registration (50% off). There are about 1000 people there, mostly semi-regulars, few students. Most of them quite technical and experienced.
  • Arrived Tuesday night, Wednesday first day of the conference for me but the conf starts on Monday with small sessions.
  • results
  • Attended talks, among others by Red Hat/Fedora ppl Learned about Koji and how it is quite a bit behind on OBS (which is probably also due to a different focus. Either way, forget about cross compiling, it runs locally, has to be controlled from the command line etc)
  • Got in contact with a few local (open)SUSE people, will try to help them to set up more of a community in AU.
  • Realized many people want to spread openSUSE but don't know how. They used to have local Novell contacts but those disappeared (layoffs, personnel changes) and now they're lost. I will try and get them motivated to become openSUSE ambassador. Need to expand our ambassador program, communicate it better and work with the local Novell/SUSE offices, integrate their work with local (open)SUSE people and the ambassador work!
  • Had a meeting with Fedora Project Lead. We discussed cross-distro collaboration. There will be a discussion panel on that on FOSDEM lead by him and the Debian project lead, I will join them.
  • Communication
  • Gave a 90 second lightning talk (24 slides about SUSE Studio) during the closing ceremony which earned me a t-shirt (winner best lightning talk). Resulted in ppl coming up to the booth all Saturday to say they would check out openSUSE and SUSE Studio (again)
  • Held a booth on Saturday together with Tim, local Novell/SUSE employee and openSUSE volunteer. Answered many questions.
  • Created 1 A4 poster with the marketing team to get geeks' interest, gave away about 40 and used it in booth area
  • Gave away 150 DVD's, rest is with 2 local openSUSE ppl now.
  • Conclusion
  • This conference is certainly worth going to but needs more marketing materials and more people, a team!