February 12, 1998

The Grace of the City

The city was kind to me today.

In many ways, it was a very fortunate day and I was taken care of by both the rain gods and the train gods in very fine fashion. I limped out to the Muni station only three blocks away and although it had been raining really hard for the previous hour, it stopped while I limped out with my coat and bag. Just as I reached the platform and got a chance to look around and ask an old Chinese lady if it was the N train and she finally comprehended what I was asking, she nodded, pointed at the on-coming train and proudly announced, "The N-train."

And we got on.

I was slightly worried about missing the Powell station, as it seemed to be more like a bus than a train, per se, so I looked at the names of the streets before the station on my little map, memorized them deeply and then well before we even got into the city the train dipped underground! Surprised the hell out of me, because I didn't remember any underground stations in San Francisco. Powell station came up soon enough, and I got out and headed up the escalator for the first level and then had to hit stairs for the second. By the time I got up to ground level, I could see that it was raining pretty hard.

But the City of San Francisco was doing some construction work along 4th, and it was upper-story work, so they had scaffolding up with plenty of space underneath for pedestrians. So I remained pretty nearly dry until the last block, which had a completely new building that had overhangs, very deep overhangs for patio dining at the various restaurants housed in the building. Talk about a nice way to go.

And for the very last half a block that included crossing a street and walking along the Moscone Center, Bob caught up with me right then and put his umbrella over me. A very keen morning, all in all.

The morning session was the condensation of a two hour talk into an hour and a half talk on advanced user interface design. It was delightful, filled with practical information, good experience stories, and some solid numbers to back up the experiences along with documented usability tests. It wasn't as entertaining as the tax interfaces talk, but it was very good about implementation details, what to avoid and what to follow up on as well as where the most bang for the buck could be gotten with a UI. He ended on a very interesting and very rushed note, that an application should, likely have two front panels, even if most of the functionality might be the same, the things is that the differences between the needs and capabilities of expert users has gotten so far beyond what most beginners need that it's going to be necessary to have two front panels to meet those needs. A lowest common denominator front panel could work five years ago, but that for 1998 it just wasn't going to work anymore.

I could believe that. Lunch was a quiet indoor affair at the restaurant over the Center, and I think for the same amount of money that people were paying for ham sandwiches I got angel hair pasta with pesto and a grilled chicken breast. Yum.

The first session in the afternoon was on business rule modeling. It was from a contractor that worked a lot with business people that didn't want to know anything about computers or applications and it showed. She was fine about defining what a business rule was, but incredibly vague and nearly coy about how to implement them into something useful. While she was able to declaratively say that business rules were underneath every part of any system and that they should both be made visible and variable by the very business folks and people that were using and changing those rules, she gave absolutely no clue as to how. Yes, they are a consulting company, and they will, for a fee, a very fine fee, come and teach you their steps and tricks for getting business rules into a workable design.

But I got three good things from her, some of them fairly primitive, like every single thing you put in an application should stem from a business rule, i.e. it has a direct impact on your business and it grows from the purpose and reason for your business to exist. The second idea was that those rules should be visible and configurable. Much like the configuration stuff we make available to our expert users, which they can use to change the rules of how the application works for them and their users. What we need to do is make those things even more usable and more configurable and more visible. Mostly by knowing what those are to begin with.

The second session was on testing in such a way as to save face and money. Most of it was stuff I already knew about testing. That it's easier to have test cases to begin with, that there are mass benefits for planning for test rather than just saying, in a lump, we need to test. One good thing is that if tests are broken down much as a design is broken down, one gets a better estimate as to how long tests will take. The other great idea was that tests shouldn't be generated from the module architechture of the system. Or not just from that. That there should be a class of tests that are derived directly from 1) what does the system have to have to work and 2) what will break the system? The two questions that are kinda orthogonal to the usual test procedure set. I liked that.

Bob and I went to dinner after that and were talking a mile a minute. That was really fun and gave me a good way to digest what had been said and put it into my networks more solidly. There were a lot of cases that were discussed in the sessions in the previous day that make for really, really great stories, and the guy doing the advanced UI talk today had some really great ones as well.

I then called Bryant's work machine to tell it that I was just going to go home on the Muni as it was close, it was dry out and I thought it would save him some time and effort. It was an utterly clear and cold night and the walk was nice if slow. Also, just as I reached the part of the platform that I could get on the train without straining my knee or myself, the N train appeared! The ride back was far less nervewracking than the ride in and it was fun. But I got back to Bryant's house about five minutes before he did. He'd actually been in the City, doing things and when his pager went off on the machine at work getting a message he went to Bob's hotel to look for me. Oops. He also checked out the restaurant and then went home after checking his messages.

He was very clever, and I was sad about foiling his fine plan.

But he made up for it beautifully by rubbing my feet. Sweet man.

Also got to talk with John for a very long time tonight and it was good to hear him. I am finding that I'm missing him and home and the like, but I'll have fun anyway... and learn a lot more tomorrow.

© 1998 by Liralen Li.

[ Previous | Index | Next ]