Tuesday evening, went out to dinner at Fuki-Sushi (yum!) with old friends Abbe Don, Mitch Yawitz, and 'Vagabond' Jim Home. Then over to PARC as a group for Terry Winograd's BayCHI talk on theory vs. practice in human computer interface. There ran into Harvey Lehtman, Jerry Morrison and other ex-Apple colleagues.

Most of the talk, and almost all the Q&A, was the common back and forth created by having fragmentary theories, partially embedded in tools and practices, and thrown into deadline projects in a bloody-minded industry. But a couple of things did stand out:

Terry's currently on a sabbatical working at Googleon UI issues. The team there has the capability to perform experimental HCI work on an unprecedented scale. For instance, they could add drop shadows on the tabs just above the search box, put up that version, divert 1% of the incoming traffic onto it for a while, and find out that it increased usage of the alternate categories by 17% with a 99.999% confidence factor, given they had then tested 100,000 interactions (numbers from memory).

We played with testing on the users during my time at CompuServe, but this scale is boggling, and revolutionary in both an HCI and business sense. On this front, and in improving their search technology, Google is learning from its customers on an unprecedented scale. If it can keep doing so. Terry mentioned his worries that taking Google public could eventually compromise its commitment to building a long term trust relations with users by (for instance) abandoning popups and intrusive banners. My immediate reaction was 'Who would be that silly, they surely understand they have a good thing going?' Then I remembered the first search engine I loved, Altavista, and the second, Excite. Both captured and essentially killed by the content/portal/sticky/convergence crowd. It could happen again. Keep fighting the good fight guys, and try to get some smart shareholders.

And then the lamest moment of the night. A comment from an unnamed audience member to the effect that 'surely we can get design right this time, on ubiquitous computing, since we can all see it coming.' Yeah, right. You tell me the form factors, target markets, and business models that win - just for starters - and then we can talk, right after you walk on water. Terry tossed that one off, but he might have mentioned that the whole theory and design issue is embedded in an evolutionary system, a market, that ultimately learns from the users, though not in a predictable path. For instance, the Newton was theorized and designed to hell and gone. Unfortunately, neither the technology or market were ready at the time. Someone else's design won. That's the way it's going to be again, in the Cambrian explosion of mutant computing and sensing devices.
3:27:27 PM    

Added John Patrick - I once served on a board with him - and Jerry Michalski, ex-editor of Release 1.0, both under Deal Flow. Also, Salam the Baghdad blogger, under Foreign Exchange. (Some say he's a CIA spoof. If so, it's a damn fine job!)
2:50:10 PM