October 13, 2005

Lili Chang (Microsoft)

Lili Chang spoke with us about future directions for the Windows OS.

One highlight of the lecture was what happened behind the scenes, so to speak. Dan Phiffer and I used SubEthaEdit, a social text-editor, together to take some collaborative notes on (and chat a bit about) what Lili was saying. He and I were the only people using the software, and we were sitting one seat apart, so the experience felt a little redundant, but I'm excited about the potential if we were to get a group of 5 or 6 (or more) collaborators together. For one, I would end up with a much better outline of the speech than I could do myself, and I think having a running commentary/outline would help me to concentrate better on what is being said.

But back to Lili Chang. She's a Tokyo native, raised in Nebraska, and was originally trained as an architect. She worked at Apple for a little while during the horrible years when John Sculley was in command. She's been at Microsoft for 10 years, and before heading up Vista, she was working in some smaller department that I can't remember.

Some of the highlights of Vista/Longhorn:
- Huge emphasis on "search."
- Visual cues more important to enhance OS experience
- Less hierarchical approach to file organization
- At the same time, Windows is trying to keep straddling the line by changing the OS in ways that allow their corporate clients to use the upgraded software with no new training.

Lili's speech was very buzzword-heavy and it grated on me that there seemed to be little actual substance. She only superficially touched on many of the subjects. For instance, she seemed unwilling or unable to discuss why "search" was so good; it was simply assumed that it would enhance the user's experience. She would only say that, "the search people are working really hard and it's going to be great." There could be some sort of non-disclosure stuff holding her back (we weren't allowed to take pictures or record video or audio), but it didn't really seem like it. She was forthcoming with other details, like their testing methods, which I also take some issue with.

According to Chang, Microsoft has beta-tested Longhorn with a couple hundred families across the U.S. We were shown how they use a special application which places a smiley and a frowney face in the toolbar that the beta tester can use to give positive or negative feedback. We were treated to a number of the candid comments users gave, which were almost entirely focused on the minutiae of the interface: Someone didn't like how a word got truncated if a column was too small, someone else liked the way the new trash icon looked, etc.

Chang was really concerned with these types of details. She told us that these little details were the most important part of the experience for the users. I disagree. I think that the visual flourishes of an OS are important, but much less so than the larger issues of continuity of user interface, intuitive-ness of the design, making sure the underlying processes run correctly, etc.

I'm also not sure that beta testing for design purposes is the best method, at least not by testing on arbitrary computer users. Beta testing is good for catching bugs, finding out how people use the system (they usually use it in ways never envisioned by the designers), and so on. But if you ask people of average design/user interface experience, you are likely to get average (i.e., mediocre) responses. Like, "I like the trash can icon." Well, it's good if the icon is appealing, but if you ask a user interface designer, you might get a reason that the trash icon works as it does.

Also, the beta testing is only going on in the U.S., according to Chang. Earlier in her talk she had mentioned the differences in how different cultures like to use their computers. But, since Longhorn hasn't been localized yet, they are only testing in the U.S. I don't think localiztion would be likely to include much more than translating the text, so it looks like Microsoft will end up with a U.S.-honed version of Windows that is going to be deployed to the world.

In retrospect, the thing that bothered me most about what Lili had to say was her almost flip attitude toward the whole thing. I was expecting a more nuanced, thoughtful, (intellectual?) approach to designing the look and feel of an OS. I think this is the crux: Chang seemed more concerned with making it look right than in actually making it right. That may be too subtle a point to pick on.

This is not to say that the focus of her talk was all on the look and design. There was a fair amount said about the technologies and ideas in place, but as I said earlier I found the discussion of these ideas mostly superficial; they didn't seem to really be part of the OS because of a considered look at the needs of the system, but more as though they were necessary almost by Microsoft fiat. Search is just inherently good, it just is, is kind of the impression that I had.

Despite these points of contention, I enjoyed hearing her speak. The design of an OS is a fascinating issue, and what she said got my gears turning. I'd really love to hear the other side of the story, maybe from a current Linux or an Apple architect.

Posted October 13, 2005 03:28 PM. Categories: Week 6 | Permalink