Issue #20 June 2006

Red Hat Speaks

Havoc Pennington, Engineer

So there's this old interview we have in the archives. Back then, we had just announced the release of Red Hat Linux 8.0. Now, we've just launched Mugshot. What's different now? What's the same?
In that interview I was talking a lot about achieving basic sanity on the Linux desktop - making things usable at all. We've made enormous progress on that since then and continue to do so.
When I sit down to think about priorities in 2006, though, it's not enough for open source projects to achieve parity with the competition - ultimately you can spend millions of dollars and years and years working on that, and it's more like an endless treadmill than a race you can win. Rough parity is important, but it's the wrong focus.
That's why you see us exploring as many new ideas as we can. One of them is Mugshot, another is One Laptop Per Child, with more in the works. We're also orienting our daily work and culture around practices we refer to as Design Thinking.
One of our goals with Mugshot is to try creating an open source project that incorporates design practices and culture, rather than hoarding these innovation tools inside the Red Hat Intranet. It's an interesting experiment.
So--Mugshot. I've tried it out. It's IRC and music sharing and GAIM all-in-one. Pretty cool. Where did the idea for Mugshot come from? How did something like Mugshot get started at Red Hat?
If you look at my attempt to summarize design thinking, one of the tools for getting unstuck is an exercise called definition. What are we really trying to accomplish?
As we wrestled with "desktop strategy" a couple years back we realized that the very name of the problem - "desktop" - was self-limiting. It didn't describe what we were really passionate about. The right definition had to do with desktop users. For Mugshot we're asking "what can we do that will be new and exciting for mainstream audiences?" and for OLPC "what can we do for kids who've never used a computer before at all?"
Once we opened up the definition of the problem, we thought about all kinds of things we could do, and we went with Mugshot because it was one of the smallest and simplest starting points - something we were able to get started and launch with just a few developers in a few months.
I've read the press on Mugshot, good and bad. Lots of people are excited about it, but others seem hesitant. I've even seen a few comments that worry work on Mugshot will delay improvements to Red Hat® Enterprise Linux® or the Fedora™ Project. Is this a valid concern?
Companies don't really work that way in my experience. The Red Hat Enterprise Linux and Fedora teams continue to grow, and I expect will keep growing as demand for Red Hat Enterprise Linux and Fedora continues to increase and our customers ask more of us.
We moved only a couple people, out of hundreds, onto the Mugshot project - and we hired replacements for them within the OS team.
So this is "in addition" not "instead."
What part does Mugshot play in Red Hat's business and commmunity offerings? How does it fit into the company's goals? Is it a play in the same space as MySpace, LiveJournal, or FaceBook? If not, what is it?
A lot of those questions aren't answered yet. We wanted to start with the idea of "mainstream," we researched some more specific groups such as "teens using MySpace" or "TV lovers," and we asked "what can we do for these people?" - that has to be the starting point, that we're providing something some specific groups might really like. (It's OK if some other specific groups hate it, btw, but someone has to like it. MySpace is a great example of something many people love and many others hate intensely.)
Our first goal is really to learn. To create Mugshot, we've had to learn relatively simple things (e.g. how to work around Internet Explorer bugs), and also much more complex things (e.g. what today's teens do with their computers).
On the topic of "is it the same as MySpace" we have a blog entry that goes into some detail on that.
I guess I'm dodging the meat of the question though - which might be "why mainstream and not enterprise?"
I view this as a return to Red Hat's roots. Red Hat became an enterprise-focused company only in the last few years; I've been here twice that long. Our goals were always larger and we want to be relevant enough to promote open source, open content, collaboration and freedom well beyond the enterprise. Right now the company has the breathing room and credibility to explore new directions and innovations and we're taking advantage of it.
Now, a few questions everybody's been asking about Mugshot:
  • How do I get a login?

    Sign up on the site for the waiting list. We have a backlog of a few thousand, but we're trying to scale up capacity as quickly as we can and inviting more and more people as we do so.

    We'd love help by the way... there are so many bugs to fix and features to add, we can't get to anywhere near all of them, so more hands will help us invite more people more quickly. We are also pretty quick to give you an account on Mugshot if you submit patches!

  • Mugshot is currently beta--when will the first full version be released?

    I don't know if we'll have a formal release; we think of the current stage as a full version of sorts, just not very scalable - so we have this invite-only limited user trial. We'll invite more and more people as we can scale it up further, and I don't know if 1000, 10000, or 100000 people would count as the full version.

    We'll also be adding more features, so we'll probably always have some features closer to the prototype stage, and some more mature.

  • Currently, Mugshot has features that let users share links and music and chat with each other. It looks like a television feature is slated for development soon. What other sorts of features are you thinking about?

    We keep a partial list on http://developer.mugshot.org/wiki/Activity_Ideas

    I'm most excited about what we haven't thought about yet, though... there's always a new idea just around the corner.

  • Mugshot is an open source project. How can interested developers, designers, writers, and users get involved?

    The best way to get involved is to just dive in; pick a project, whether it's a research project or a coding project, start working on it, and share your work with the rest of the Mugshot community.

    We tried to put all kinds of information up on http://developer.mugshot.org.

What's this 'design thinking' business all about? Isn't it just for marketing people and C-level execs? The Mugshot wiki talks about it a lot. How was this process used to develop Mugshot?
Marketing and C-level execs have more trouble than most with design thinking, in my experience.
People talk about it in lots of ways, some more buzzword-heavy than others (you can blame me for most of the design-related text on the Mugshot site), but it's very real and very practical when it comes down to it.
In school everyone learns to evaluate ideas; you take a set of existing concepts or an existing argument, and take it apart. You think about corner cases a lot, and build abstractions and generalizations.
With design thinking we want to encourage people to generate ideas; whether by redefining a problem, brainstorming new approaches, making concrete prototypes rather than having abstract debates, or researching people and their needs rather than looking at existing technologies.
It's a different set of skills and a different frame of mind, and in the end you have to learn by doing, just as you learn anything else.
With Mugshot we've been practicing and sharpening our design skills. To bring newcomers up to speed we have a lot of blog entries and wiki pages about specific prototypes, research projects, brainstorms, and other design work during Mugshot's short history.
I strongly believe that skipping this work and jumping straight to coding results in substandard user experiences.
What do you do in your spare time? Do you even have spare time?
When not working on Mugshot, I am mostly about TV, movies, books, and video games; not the most admirable use of time I guess, but it explains a lot about Mugshot!
How was the Summit? What was the best party this year? Who gave the best talk?
The keynotes from Cory Doctorow, Eben Moglen, and Nicholas Negroponte stole the show if you ask me. The summit includes nuts-and-bolts engineering sessions but I love the "big ideas" talks.
On the party front, I missed most of them since we were so swamped either preparing for the Mugshot launch or all the work afterward. The Wildhorse Saloon was fun but we were hiding upstairs well away from the line dancing.
If your bosses came to you and said you could have all the staff and resources you needed, in any amount, what project would you undertake? Why?
I'd focus on quality of hiring and innovation-oriented culture, and try to start as many small projects with great people as possible. We already have several project ideas queued up that are loosely related to Mugshot, but with great hiring, my experience is that new hires quickly have better ideas than I do.