The evening began with the advertised "Special Holiday Surprise", which turned out to be 37 minutes of having holiday songs sung at us by a group called "Sunday's Child", punctuated by two or three occasions between songs where administrative announcements were made and we thought that the more substantial content was about to begin. Don't get me wrong, the singing was good...maybe it was even excellent, I wouldn't know. But 10 minutes of it would have been even better.
Richard Anderson presented "TOP TEN BAD ATTITUDES ABOUT USER INTERFACE DESIGN" (don't have the original citation; I don't think it was Richard's list):
10. "Trust me, *I know* what the user wants. I used a spreadsheet once."
9. "This menu item doesn't really fit anywhere, so I'll just put it under Utilities."
8. "This tool must be powerful...it comes with six manuals!"
7. "The standard Motif widget set is for programmers who aren't smart enough to write their own widgets."
6. "If the user can't figure it out, he doesn't deserve to use that feature."
5. "If it crashes a few times, the user will learn not to use that feature."
4. "Prototyping is for wimps. We'll get it right the first time."
3. "Code it now. We'll figure out our Software Configuration Mgmt. process later."
2. "Don't worry about the user interface now, there are too many bugs. It will be fine by the time it ships."
1. "The purpose of design is to weed out Psych majors."
For the main panel discussion, Richard Anderson presented the other three panelists with a series of issues, and let them say whatever they wanted. The panelists knew much more about each other than we did, and so came prepared to disagree thoroughly.
Don Norman: I'm here to take a contrary view. Yes, of course people who "need" to buy into user-centered design have a different view of what it is.
Janice Rohn: Popele don't have all the information they need. They aren't exposed to HCI topics during engineering/other educations. They are only exposed to value systems such as return-on-investment and cycle time reduction.
Daniel Rosenberg: Management understands only two things: Fear (of failure), and Greed.
Don Norman: "Become credible."
Janice Rohn: Credibility attacks are either (1) unwarranted, or (2) obfuscation. Get to the real reason for the resistance. Cite studies, have data ready. At Sun, "Usability Engineers" carefully position themselves as serious engineers with a particular skillset.
Dan Rosenberg: Dan talked to an Adobe executive who said, "There are all sorts of useless people in the world. ... Mothers-in-law, and then UI Designers. Including you, Dan! If you ever worked for me, I'd fire your ass." Seeing "UI Designer" on a technical resume is just as good (just as useless) as seeing "Branch Manager of a bank".
(DanR) If the medical profession was like this, with no certification, half of you would be dead.
Dan's advice: If you want credibility, GET A BIGGER OFFICE. Really. Schedule meetings with people in a big corner office, so they know you are high up in the food chain, and you can bite them. Dan has never heard a UI person described as "lightweight", only that they are annoying, or will slow down the project.
(DanR) "I don't think people care about process, unless they work at HP. People care about results."
(DanR) "When (Borland CEO) Philipe Kahn said 'I can do it myself', that got the HCI folks a lot of business. The project managers said 'We'll do anything you want, just get Philipe off our backs." But it doesn't work with Larry Ellison.
Norman advice: Make engineers watch real customers using the product. If engineering considers the users stupid, that's their challenge: Write it so that even these users can use it.
(DanR) The Sun Server group engineers are great to work with, because they are Unix Kernel programmers and THEY KNOW IT. They don't try to do the UI themselves, they just follow the UI spec faithfully.
Norman: The difference is, when HCI engineers are wrong, they're ashamed. When engineering gets it wrong, they say, "Oh well, we'll do it again."
Norman: Give UI bugs the same priority as other (crashing / performance) bugs. Usability/reliability bugs are legitimate bugs, which will cost money in service calls & support.
(DanR) At Oracle he (I think it was him) has "UI Rabbi" on his business card. He acknowledges the "religious" aspect of UI design, and claims the right to interpret scripture.
Norman quoting DanR: You get three changes. If you're wrong the first time, and you're wrong the second time, then the 3rd time you're dead even if you get it right.
(DanR) Deserves parity w/ profitability at the executive level. It will cost you $100,000 if your project did not win PC Week Editor's Choice.
(DanR) Some project won the PC Week Editor's Choice, and Dan got a phone call from Larry Ellison saying, "This is the FIRST time I'ver ever seen 'Oracle' and "Usability' in the same sentence when it wasn't a lawsuit."
(DanR) ...something about taking 10 million lines of old crusty code and trying to improve it, is like "trying to put lipstick on a pig."
(Norman) The hardest thing to change at HP is the culture. Maybe by 2001/2002.
(Rohn) Will be convinced that execs really care when they make middle management see the light. Must look at lifecycle costs, cost of ownership, not just development costs.
(DanR) What they can do is 'Show me the money.'
(Norman) It's hard to do long-term investment in HI when the stock market requires quarterly reports.
(Norman) "Live with it. It happens to everybody."
(DanR) "All is forgiven if you're winning."
(DanR) Showing a prototype of a shiny new product is very bad for sales of last year's model. Marketing guys hate that.
(Norman) Team up with marketing, not Us vs. Them. Piggyback on marketing credibility. CM, MIT, and Michigan have HCI in business school.
$12 billion out of $300 billion is sold to users. The other $288 billion is is to "buyers."
(DanR) "If you don't design for the BUYER, you're out of business." There is a difference between a business decision and a design decision.
(Norman) Companies have toomake sure you're unhappy every year, to sell you more software. He wants to throw out the PC and to Information Appliances.
(DanR) Make the product smarter, not larger.
(Norman) ...or make it simpler. There are too many details. You have to see the larger picture.
DanR & Norman: Sometimes "no human interface at all" is the right answer.
(DanR) "The Internet is the WORSE thing to happen for UI design. Hack and Throw is the only methodology."
(DanR) On January 1, 2000 we will see the biggest change in the relationship between industry and society. We'll fix all the COBOL code, but there will be no electricity. Every microcontroller in every building will turn on the fire alarm and the sprinkler system. People will take a fresh look at the relationship between software and society."
(Rohn) Sun has entered the Consumer market with Java.
(Norman) No, you haven't. Embedding in consumer devices is nothe same; you don't advertise in Parade magazine.
(Rohn) Commodity companies are better at demand creation than computer companies.
(DanR) Software that HAS TO WORK is NOT revised every 12 months. It's a 3-year cycle, with a 1-year beta period. These systems have less than 2 minutes per year of downtime.
(DanR) You need to be empowered to cancel projects and start over.
(DanR) "Shipping is the #1 feature."
Norman's interpretation of DanR's advice: "Work for a company that is making a lot of money, and be an executive." -- DanR's response: "What Porsche are YOU driving, Don?" -- DonN: "911".
DanR's opinion on Unix usability: "It's like asking a blind person what they think of painting."
Final words from Norman: "The field has to grow up. Needs more effective methods. We don't want to be a 'service team' that gets called in to help fix up bad designs. Need allies, such as marketing."
(DanR) "If you don't know the difference between accrual/cash accounting,
you shouldn't be an HCI manager."
[ Page last revised 7-Jul-2001. ]
[ www.lyons42.com | email Dave ]