Safe as Milk

39:04

Improved on my 10K personal best today. Happy, but not delighted. 38:59 would have looked so much better. Especially disappointed to have been under 3:50 for 4 kms, and over 4:00 for 4 others (not consecutive). Regular is better.

Also, I couldn’t resist having a couple of beers yesterday to celebrate Ireland unconvincingly beating England (and I never thought I’d see the day I got to say that) yesterday in Croke Park, 14-13. That probably played a role too.

Libre Graphics Meeting 2009: community fundraising campaign

After the successful community fundraising campaign last year, the Libre Graphics Meeting will be back in Montreal again this year, and once again the organisers need your help to make the event a success.

In fact, this year we need your help more than ever. With the crisis/recession/slow-down/depression upon us, companies are puckering up tighter than a snare drum (to borrow a quote from the Shawshank Redemption) when asked for sponsorship this year.

Support the Libre Graphics Meeting 2009

It’s true, the Libre Graphics Meeting lives on the margins of the business cases people can make right now. There are not millions to be made in appealing to graphic designers working on Linux. And yet, as a community event driving co-operation and development of creative applications on Linux, the Libre Graphics Meeting is valuable, and unique to the free software community.

We need the help of the people to whom it’s valuable to make it happen. And that’s why we’re asking the community to give what they can to support the conference. To allow us to bring people together to share ideas and hack on code. To make better Free graphics software for everyone.

List of Free Software non-profits

Michael Dexter of Linux Fund spent some time last week putting together a list of non-profit corporations which exist to support free software, free software projects, or free culture worldwide. It’s an impressive list of 68 non-profits from around the world, with varied goals and focuses, but with the common theme of wanting to improve the world through the sharing of information. He published the list on the FLOSS Foundations site, which is appropriate, since that group brings together representatives of almost all of these organisations, sharing information useful to everyone.

Some of them have, in my opinion, a more tenuous link than others to projects – such as regional satellites of the FSF for example, or the organisers of SCALE – but the list does show the breadth & depth of the domains where free software has made an impact in the past couple of decades.

Maemo Community Council referenda: the results are in

The voting is now closed in the Maemo Community Council referenda on election procedures for the upcoming and future Maemo Community Council elections.

The provisional results, which will stand unless successfully challenged within the next 7 days, are as follows:

1. Which of the following criteria do you want applied to the Maemo Community Council elections to be held in March 2009?
   • 25 karma and 3 month old account (status quo) (71 votes)
   • No karma requirement, anyone with an account for more than 3 months
     may vote (64 votes)
   • None of the above (4 votes)

   139 votes were cast of an electorate of 601 (23%)

   We will maintain a 25 karma requirement for the next elections.

2. Which of the following criteria do you want applied to future Maemo Community Council elections?
   • 10 karma and 3 month old account (48 votes)
   • 25 karma and 3 month old account (status quo) (46 votes)
   • 10 karma or 12 month old account (22 votes)
   • No karma or account age requirement – everyone with a maemo.org account may vote (13 votes)
   • None of the above (0 votes)

   129 votes cast of an electorate of 601 (21.5%).

   For future elections, voters must have a karma of 10, and have created their maemo.org account more than 3 months before the closing date of the election.

3. Which of the following voting systems do you want used for Maemo Community Council elections?
   • A single transferable vote preferential system (60 votes)
   • No change – single vote, with top 5 candidates elected (51 votes)
   • A reweighted range voting system (score candidates between 0 and 100) (13 votes)
   • None of the above (3 votes)

   127 votes cast of an electorate of 601 (21%).

   Future elections, including the election to be held in a few weeks, will be by preferential vote, and will be counted using the Single Transferrable Vote system.

I will be modifying the Maemo election system (stolen borrowed from the GNOME Foundation) before the next elections to implement preferential voting, and we will likely be using OpenSTV to count ballots after the next election.

Governance best practices

Jono asked on the AOC blog for successful governance stories, and while I’m happy to comment on the blog, now that I’ve taken the time to write some down, I thought I might as well share them 🙂

Governance comes in many shapes & sizes of course. My favourite governance stories are about federating individuals, who manage to channel community efforts, maintain a meritocracy where code talks, and yet don’t come across as authoritarians.

Outside of Linus (who’s a good example), Ton Roosendaal of Blender has this kind of presence. Talking to Ton, it is easy to see that he cares about Blender and about the Blender Community. The care and attention that he brings to projects like “Elephants Dream” and “Big Buck Bunny”, or to the supporting documentation and conferences he organises for the community, illustrate the esteem in which he holds his users and his developer community. Even the way the Blender Foundation came into being was amazing.

One of my favourite communities is Inkscape. When they broke from Sodipodi, there was this acrimonious flame war, and something of a bitter taste in people’s mouth. So what Bryce Harrington, Nathan Hurst, MenTaLguY and Ted Gould did when they split was decide to throw open the doors, and accept code from all comers. They set a direction and some ambitious goals, but they were very clear from the start – come right in, you’re welcome. And this gave the project some great results, especially early on when it was still establishing itself. Bryce describes one of them in this article.

The success of the Inkscape project’s governance model is borne out by its ability to escape founder’s syndrome – Bryce, Nathan and Ted have now backed away from the project to some extent, they’re still there as wise heads, but they have passed off the direction of the project to other capable people.

I think the way that Drizzle was born bears some resemblance to this, and I really like the way they have consciously broken down the walls which were necessarily up around MySQL. Brian Aker’s been something of an inspiration on this. His mission statement at the announcement of the project was astounding.

Subversion‘s governance model is an exemplar of best practices too. Set a clear project scope (“Subversion is a compelling alternative to CVS”), clear goals, establish transparent and fair community processes, and open up the gates. Anything within the scope of the project is fair game. And once again, code talks. This story, from Karl Fogel’s “Producing OSS” illustrated the robustness of their governance, and the confidence the project’s leaders had in their ability to influence the project.

The Maemo Community Council has the potential to be a very good governance structure, I think.  The idea of a governing body of the community, by the community, for the community, whose goal is to canalise the efforts of a disparate group into something coherent, and to provide a legitimate point of contact for technical decision-makers in Nokia, is a novel one, and hasn’t been tried, as far as I can tell, by other companies.

Counter-examples of good governance are all around, I won’t name any in particular to protect the guilty. But many of them stem from a misguided belief in absolute free speech, to the detriment of the quality of discourse and code in the project (“we are all created equal”) which results in very chatty, but unproductive, individuals taking senior positions in the community, or a sort of shyness of the founder or leader, who doesn’t believe that it’s his place to set a direction and tone.In company-run projects, excessive control or influence is an equally toxic characteristic. Companies who retain a veto on community decisions are companies who do not trust their communities.

Think you should have a vote?

For those of you who believe that you should have a vote, but have not yet received ballots for the Maemo community council referenda please let me know. A reminder of the criteria for eligibility: all those who have had a maemo.org account for longer than 3 months from the election end date (the account was created on or before the 23rd of November 2008), and who had 25 karma or more on Sunday evening, have a vote. That makes 600 of you.

Hopefully, there is no-one in this situation, but if you feel that you should have a ballot, please drop me a line with your maemo.org account name, and I’ll look into it.

Maemo Community Council referenda: voting now open!

Voting is open for the 3 referenda around the election procedure for Maemo community council elections.

A wiki page has been created for each referendum, where arguments for & against each of the options can be listed. For the moment, these pages are skeletons, but feel free to debate in the Talk page, and perhaps add some arguments for & against in the main page (but they will be heavily moderated to avoid flame-wars).

A reminder of the 3 referenda:

1. Election procedure for the next Maemo community council elections: The choices are the status quo in the election procedure document (25 karma + 3 month), or removing all requirements. The council is recommending a vote to remove requirements.
2. Election procedure for future Maemo community council elections: The choices are the status quo, lowering the karma requirement to 10, and maintaining a 3 month requirement for accounts, lowering the requirement to 10 *or* 1 year since account creation, whichever comes first, or removing all requirements of karma and account age. The council is recommending lowering the karma requirement to 10, and maintaining a 3 month account creation limit.
3. Counting method: Three choices are proposed: first past the post, preferential voting using single transferrable vote, or reweighted range voting.

Voting is open now, until midnight UTC next Monday, the 23rd of February. As we used to say in Ireland in the ’80s, vote early, vote often!

Migrating to Linux: Tip 1: Do DHCP and DNS first

If, like most organisations, your client PCs get their IP addresses and DNS and routing information over DHCP, you can make your migration life easier by moving your DHCP server first, followed by your local nameserver.

Whenever you need to declare a server name or IP address for a service (mail, time, proxy, whatever), use a service-specific domain name: smtp.mycompany.office, imap.mycompany.office, ntp.mycompany.office and so on.

Since you control the DNS server, you can incrementally move these over to your Linux server and you only have to change the IP address once, in your central nameserver, rather than do the rounds of all the clients every time you switch a service, allowing a painless incremental approach to moving basic internet services which, if you do things right, people won’t even notice.

Next: the hard stuff: Exchange and Outlook.

Edit: Update .local to .office – .local is reserved for mDNS (thanks to my readers for spotting this!)

Server migration: the easy stuff

In gathering material for my series on migrating to free software, one thing immediately jumps out at me.

If your server software uses industry standard protocols to communicate with your client software, then finding free replacement software is easy, painless and transparent for the user. Need a DNS service? Bind’ll do, thank you very much. SMTP? You’re spoiled for choice – there’s  Qmail, Postfix, sendmail among others. IMAP, POP3 – try Dovecot, or the UWash IMAP server.  SSH – OpenSSH. FTP – PureFTPd, VSFTPd, proftpd are all fine.  HTTP – Apache os one of many web servers available.

Pretty much anything with an RFC has free software implementations that are complete, and compare well with commercial competitors. Often, as is the case of Bind or Apache, they are the leaders in their space.

In other words, by using only standard client-server protocols, you have freedom to leave.

However, if your server software “integrates” tightly with your client software, in the style of Notes and Domino from Lotus/IBM, or Exchange Server and Outlook, or Sharepoint and Office, or if it has its own proprietary wire protocol, then you may have a pain point.

So the first lesson, I think, is consider how replacable server elements of your infrastructure are at the acquisition, if you want to avoid lock-in later on. As hard as projects like Samba and Zimbra chase the tail-lights of proprietary wire protocols, the easiest way to avoid them is to rely, where possible, of standard, open protocols.

And that’s what I’m looking for more than anything. How do people get around their pain points? Have people had an Exchange or Sharepoint hang-over?  Now that PostPath has gone away, are people looking to get rid of Exchange stuck with Zimbra? Has migrating from MS SQL to a free database server been a pain in the leg? What have people used to centralise authentication and share home directories across the network? Is Samba with LDAP a drop-in solution?

Looking for case studies

I hope to be writing a series of articles over the coming months on migrating from being a Windows shop to a free software shop, broken down stage by stage:

1. Replacing Windows file sharing, web servers and services and email servers by Linux + samba/apache/postfix/bind. Move proprietary applications (SAP, Oracle) which have Linux versions to the new Linux servers. Migrate desktop software to open standards provided by these services (break the Outlook + Exchange habit).
2. Replace proprietary desktop applications by free software equivalents (low-hanging fruit: Firefox, Thunderbird, OpenOffice.org)
3. Require use of open formats for company documents where practical (move from .doc to .odf)
4. Identify desktops with no windows-specific business applications installed, and migrate them to Linux (thin-client, perhaps)
5. Move from proprietary business applications to web-deployed free alternatives (Alfresco, SugarCRM, …) and start redeveloping vertical apps as web applications
6. Migrate commercial server applications to free equivalents, where feasible
7. Migrate remaining desktops & laptops to Linux (with some possible exceptions for task-specific desktops, like graphic design, accounting or CAD?)

My goal is to find people who have braved this storm, and have gone through the stages of migrating from windows to Linux, who can share the pain of each stage, and who can provide insights on making things easy for those coming after.

If you are one of these people, or think you might know one, or you’re just interested in telling me how (great|horrible) the idea is, give me a shout in the comments, or drop me an email at my gnome.org address: dneary.