Archive for October, 2005

Understanding the Dapper Drake Development Process

My first day at the Ubuntu Developers Conference (UBZ) has been a lot of work and a lot of fun.

There are a few key decisions going into the next release cycle that inform everything else.

As I mentioned yesterday the next version of Ubuntu is going to be supported for a long time. In order to make that happen, the Ubuntu Distribution team is going to make a couple of minor changes to the release process. Most of these come under the basic theme of minimizing large scale changes.

Earlier on in the discussion of the Dapper process Jeff Waugh suggested the possibility of not syncing up to the latest stuff in Debian unstable on this cycle, and just moving forward from 5.11 base. The idea behind this suggestion was to limit the number of unknown bugs that would be added to the distribution. However, Matt Zimmerman pointed out that ” Debian and upstream fix more bugs than we do, hands down” so this idea has been rethought a bit.

That said, there is going to be an overall reduction in the number of new crazy things going into Ubuntu. But as of last night there will be one major exception — the LiveCD based installer.

This is making it in, not just because it will make the user experience better, but also because with a LiveCD installer Ubuntu will only have to ship one CD through ship-it. And that in turn means they will be able to support more than one flavor of Ubuntu. This means that they will be shipping Kubuntu cd’s for all the KDE centric folks out there.

But all of that comes from this morning’s first talk, and since then I have been participating mostly in talks about Ubuntu Server stuff, and I’ll try to blog about that more soon.

Dapper Drake

Dapper Drake is the code name for the next version of Ubuntu Linux. It will not just be supported for the standard 18 months, but for 5 years on the server, and 3 years on the desktop.

This is possible because we are looking at a confluence of upstream stability. GCC,, the kernel, and other key components are now at, or are quickly approaching stable releases, so it only makes sense to capitalize on this to produce a long-term stable release.

This is not a move away from the 6 month release cycle that Ubuntu is now famous for. So, there will be 10 new Ubuntu versions before Dapper Drake is no longer supported.

It will be interesting to see what this does for Ubuntu in the enterprise. I’m sure that some large enterprises want this, but I’m also pretty sure that they are going to want a few other things too.

I will have more updates on stuff that was discussed today this evening.

Ubuntu Below Zero

I’ll be flying out to Montreal on Saturday to attend Ubuntu Below Zero. I’m a serious fan of Ubuntu Linux, and of the people that make it happen. And I’m excited to start participating more in the whole Ubuntu community.

So far my contributions have been in giving talks on Ubuntu at LUGs and conferences, helping install Ubuntu on a lot of people’s machines at various install fests, reading the Ubuntu mailing lists, and assuring questions for the folks around Ann Arbor who have Ubuntu related questions. But I am really looking forward to crossing over the line from consumer/advocate/helper of a pre-existing product to someone who participates in the development of the new version.

I want to make sure that it is drop-dead easy for developers working on Rails, TurboGears, Cake, or any of the next generation of web development frameworks to live and breathe Ubuntu.

I also want to make it easy to create a meta-package with the tools that a small business needs to get up and running on Linux quickly and easily. We have all the pieces in place (well, calendaring needs a bit of work, but it looks like it’s going to be there soon!). Unfortunately, there’s so many choices and the configuration and integration project still too hard.

A lot of small businesses set up Microsoft SMB Server 2000 as their first server. It’s cheap and easy and it works — mostly! — and then when they grew out of it, they are stuck paying thousands of dollars for upgrades to the “real” Microsoft back office products. And for a small company all that money that could be more profitably spent in developing their core business.

All of this resolves down to my key goal — I want to reduce the cost of entry for small business, and make it easier for new companies to get started without the need to go to venture capital firms. Ubuntu, Linus, and open source generally can reduce the need for significant capital expenses early in the life of a small business, and in the end that is going to improve the economic outlook for small businesses, and help the economy, and facilitate innovation.

Lean Manufacturing Practices — Kaizen

When I was a kid my mom had something called a white tornado — we would all get together, and clean for 20-30 min. In the end the whole house would look better, but equally importantly we all felt like we participated, so we felt more responsible for keeping it that way.

At Toyota they have been doing the same thing for decades. The Lean Manufacturing world calls it a “Kaizen event,” and it is a key element of the process which Toyota have used to build cars faster, cheaper, and better every year. They gather together a team and ask them to spend a day working to come up with improvements, and then go out on the floor and implement them.

The Kaizen event’s first and most obvious benefit is that provides time for improvements to be implemented immediately and regularly. But when it is a regular process, there are other benefits. Employees are encouraged to develop a pattern of looking for improvement opportunities, so they can have something to suggest to the team at the next event. The whole team takes ownership of the improvement process, and individual employees begin to see themselves as a team. There’s also a certain energy that is generated by making changes in real time.

At Toyota Kaizen is not just an event, but a continuous process of improvement, but the event is a key to developing a culture where continuous improvement is just part of the routine.

IT Departments need Kaizen events, new technology is coming at them faster than they can manage, processes aren’t automated just because nobody has a free couple of hours, and things can get messy very quickly. If your people aren’t motivated to improve the processes in their area, and feel frustrated with the demands of their work, you can help them by carving out time to have a Kaizen event once a month. Give them the tools and resources they need, and let them loose improving the things they touch every day.

(Almost) Free class on Ruby on Rails

I will be hosting a free Ruby on Rails “tutorial and hack session” in mid to late November. This will be an intensive session for those who want to get up to speed on Rails quickly. We will learn the basics of Rails, and do some actual work on an Open Source app, so learning will be balanced by doing, and people should leave the session feeling pretty comfortable that they can move forward on their own.

Tentatively it will be from 10 am to 10 pm on the Saturday before Thanksgiving. I think 12 hours is long enough to get something significant done, but short enough that it won’t be totally mind-blowing. We’ll order out for lunch and dinner, so it would help if people could throw a few bucks into the pot to help me defray those costs.

But other than chipping in for food, and doing some work on an open source app, the whole thing won’t cost you a penny — at least for those who can bring their own computer. For those who can’t I believe we will be able to provide pre-set-up computers for you to use, but there will be a small fee to defray the costs of providing that service.


Update: There seems to be enough interest, and I’ve found a place for us to meet, so the date is confirmed! We’ll be meeting November 19th, from 10 am untill about 10 pm.