Converted to HTML by Mike Lee 3.26.2003
Updated 3.27.2003: Corrected URL and name for Magoolaghan, added a couple more links to Wayfinding Panel

Full conference schedule on ASIST site:

http://www.asist-events.org/IASummit2003/schedule.shtml

Saturday, March 22, 2003

:: Keynote: Stewart Brand

Doesn't want to be thought of as an architect because "architects design and then they go away."

When thinking about architecture (either physical or information), different aspects of the structure change at different rates. That has be to be acknowledged in order for the structure to effectively grow over time. Stylistic changes, in particular, happen all the time and a structure that doesn't allow for that will ultimately not function over time. Called "pace layering".

"The loathing of architects usually comes from the people who live in the buildings." He implies that "signature buildings" (buildings as a piece of signature art such as Frank Gehry) are inappropriate since they don't adapt. These buildings "die" when they go out of style.

Builder John Abrams (Cape Cod) takes photographs of the walls of the houses he builds just before the walls are closed up. Then this is put into a manual with specs on everything used in the house and contact information of everyone worked on the house. The manual is handed over to the owners and then the owners hand it over when the house is sold.

The first solution is not always the best solution. The fact that you have a solution that works is no excuse to stop designing. The next solution (tweak) you come up with might be better.

The lack of continuity with digital information (web sites) means we are creating an amnesiac society -- which is inherently dangerous.

The Long Now Foundation (Brand’s latest project) seems to be about avoiding the amnesiac tendency. The Rosetta Project has an interesting comparative word generator the goes across multiple languages.

He is quite the pure democrat. He likes to see meta trends by polling or comparing across culture/time. He calls it "emergence" or "gardening" a culture as a basis for determining the best design.

Why does he work on the slow (large, global issues that take place over 10,000 years) instead of the fast? "If no one is working on it, you can have a monopoly and have a pretty profound affect."

The US Constitution is an example of a document that is "inhabited". A document that was uniquely designed to be flexible over time and functioned well. Perhaps more web sites have constitutions: a set of rules of how the site functions and how it was built. It would have to allow people to participate in it, but the set of rules should be adaptive enough to absorb and learn from differences.

Suggests that since computers don't work, we (as a society) have developed the art of the work-around. We don't expect things to work the first time. We are quick and creative about finding other ways to get things done. [I wonder if that will change as time goes on and computers get more reliable. Are we living in a unique time of creativity because of this obstacle?]

How to keep something from going bad? Build in ways for the system to tell you that something is breaking. Put something that smells bad when it gets wet inside your walls. Natural gas is an example of this.

Stewart Brand is an excellent speaker. He’d be a good poster boy for AARP because he demonstrates against the stereotype of the older adult. He is an old man who lives very much in the future.

**************************************************

:: Wayfinding Panel

Participants: Mark Bernstein, Susan Campbell and Andrew Dillon. Moderator: Rashmi Sinha

Susan Campbell has a lot of background in testing as well as architecture. From the Bay Area.

He says that the “footnote paper” (an early work on the topic that everyone else refers to) he did on breadcrumbs was a waste. He found when he actually tested users on hypertext, that they never actually felt lost to start with.

Hierarchy isn't the answer and search is even worse. Search is the last resort of a reader. Context is all that matters.

Metaphors are just wrong. Don't let the structure express itself. You don't want to put a simulation up. Just build a web site and don’t force it into a metaphor of a small town or physical room. Don't make it into something it isn’t.

Spatial hypertext is inherently more flexible than hierarchical or other structure.

Tinderbox is a hypertext visualization tool he’s working on (Mac only): http://www.eastgate.com/Tinderbox

Patterns work better than prescriptions. Life is too complicated to put into a structure and fix by labeling. Cycles will uncover the more natural patterns.

"Everything is intertwingled."

Bottom line: Don't underestimate people's abilities to deal with this stuff. A person might be pretty dumb, but people are pretty smart.

Spatial metaphors can work (opposite of Mark).

Working architect as well as a working information architect.

Architectural metaphor should be used: Architecture is not about the skin of the building, but what is underneath -- what supports it.

Links how human memory works in relation to shared physical spaces. It is logical to take advantage of this human tendency.

Usability set up a measurement of success of the structure.

William J. Mitchell wrote The Logic of Architecture: Design, Computation, and Cognition in 1990.

http://www.amazon.com/exec/obidos/tg/detail/-/0262631164/

Shows a visual vocabulary of how to chunk together like tasks, needs, purposes.

Frank D.K. Ching: Architecture: Form, Space and Order from 1979

http://www.amazon.com/exec/obidos/tg/detail/-/0471286168/

-- visualizations of relationships between form and space.

Navigation is the wrong driver. He is on a site for meaning, not to navigate. If he is focused on navigating, he is getting focused on the structure rather than the meaning.

You can't get away from metaphors, he just thinks that we're too obsessed with them.

"Metaphors are like sex. talking about them makes everyone a little uncomfortable. They all think that everyone else is 'getting it'."

We are missing something inherently human about space. Meaning, imagery and value is more important. We're obsessed by findability . This is a problem with usability as well.

Information has an inherent shape. When you see it, you know when it is right or wrong.

Think of applying navigation to the inherent quality of storytelling -- how to tell a good story. It's artificial.

When anyone is working with an information space, we need to focus on the semantic side rather than the spatial processing. People create a visual model of the space in their minds. Semantics is a more powerful way to support this individualized working model.

Comment from the audience

Someone suggests that unstructured navigation through an information space is age-based. Younger people would be more open to it. Andrew says we shouldn't cling to old models just because they worked in print.

Susan is TOTALLY outnumbered. Everyone is buying into this less structured approach. This is interesting. Is this an indication that working IAs are seeing that traditional structured methods aren't working? Audience is saying "Yeah, that's cool, but how do we actually do this?"

Susan: Architecture is a start point. Not an end point. Design is always cross-disciplinary.

Stuart from UB suggests that we don't use semantics to organize information because people aren't being taught literary education -- they aren’t being taught to read and use semantics in a sophisticated way. Andrew disagrees. He says stop blaming the kids.

Oh. Now Mark is backing off. "All we really have are two dimensions. We know we have 3 dimensions to work with, but all we really deal with in our minds are 2 dimensions."

Andrew strongly disagrees. He says just because it's not easy to visualize, doesn't mean it doesn't exist.

Jared Spool. Having spend a lot of time watching a lot of people, he came to the conclusion that the issue of navigation is irrelevant to the people who are using the computer. What we aren't doing is discussing what the needs of the users actually are. Andrew agree saying that information design discussions need to be raised to the level of human meaning instead of usability.

Basically everyone is saying, "form should follow function".

To use a building metaphor. We don't need hallways and we don't need elevators. We just need links. (Eric - a speaker) Spatial modules scare him.

Andrew suggests findability is bad and we are obsessed with it. Moderator says what's wrong with studying and trying to make findability better? It's a useful way to communicate with people who are outside of the industry.

In the digital space, if there is no form, what is the function? Not answered.

Audience member suggests we look not at Wall Street as a metaphor, but very old human spaces like Rome, Athens, etc. , where wayfinding is done semantically. This is an interesting idea. How is the navigation of old European cities semantic?

***************

:: Content Modeling and Controlled Vocabularies for English Region WIL Sites

http://www.asist-events.org/IASummit2003/cs16.shtml

Margaret Hanley - BBC

Library Science person.

Goal is to create content objects that can be shared between multiple web sites and multiple CMS’s.

Had lots of people creating content - and they were geographically regionalized. Most were working on Dreamweaver and were "spending a lot of time fiddling with tables" rather than focusing on content development.

1. Improve quality of content (nav, editorial, design)

2. Harmonize content within and across geographical regions

3. Access archived content more easily (once links were broken, they just left the information on). There were good images that weren't being used.

4. Identifying outstanding content on the little regional sites. Sharing this stuff across sites. (internal and external syndication)

Content objects - smallest piece of content in the CMS. Needs to create them so that they are useful, but not so many there is an overwhelming number.

Content collections – Example: CD Review includes an article, weblink, audio clip, image.

Intrinsic metadata - Data extracted from content object (headline, byline, size of image, etc.)

Administrative metadata - author, last changed, archiving status

Descriptive metadata - The "aboutness" of the content object. (includes language and audience)

Her site audit was a discussion paper looking at the sites to understand the context of the content.

Found that the existing metadata was classic, vague metadata for external search engines. She had to do a lot of education about how to do appropriate metadata with the authors -- and redo everything

The content audit was made up of the content object library item for every content object. The library item describes each content option in detail. The content collections were also defined and then templates were created from them.

Then looked at any existing controlled vocabularies

Content modeling process was fun, but of no intrinsic use to her.

:: Wachovia Intranet

[Mike has notes coming … ]

extremely similar migration process as ours except the are totally hands off on editorial.

:: Wachovia.com Redesign Project

http://www.asist-events.org/IASummit2003/cs25.shtml

[Came in at the end of this session. Mike has notes that will be linked off the blog when ready.]

Librarian who loves controlled vocabulary.

Her problem was having to combine two extensive, mature, and carefully branded web sites into one as Wachovia Bank merged with First Union Bank

At what point in the IA development process should the homepage be designed? In theory, she likes to wait until the end of the process. The home page gets designed last. The reality is that the home page look and feel is the first thing people want to talk about.

She was lucky in that the leadership of the bank was very bought into user centered design. This was a critical component in what success she had

Different home page options:

Sifter - users self-select based on type of customers

Portal - as many options as possible

Emphasize one navigational approach over the other.

Leave it as it was (Wachovia’s tabbed approach) and just worry about the branding.

Usability testing: Men all liked the sifter option and women chose the portal option.

Dramatically consistent 50/50 response. Other than this, however, she had a problem because usability results weren't dramatic. She couldn't say "go this way, the customers have spoken". Everything was open to interpretation.

For this reason, the home page design decisions were made internally based on corporate goals, requirements, and politics.

:: Scent of a Web Page

http://www.asist-events.org/IASummit2003/scent.shtml

Joking that mark Bernstein nullified his talk this morning.

“Scent” is created by the content in the site and amplified by the links. Scent means that content moves from general to specific by way of trigger words

Whenever a user runs across an error page, then the IA has failed. Using the "back" button is like a dog who has lost a trail going back to the last place he had the train and trying to pick it up again

The chances of success drop significantly when the user hits a dead-end page. It is a predictor of failure.

“Pogosticking” is going from the home page to a content page to the home page to a content page. This is also a good predictor of failure. (Others argued with him later saying that this was simply an indication of browsing behavior and that people might be doing this on purpose.)

Use of search is also a predictor of failure. Suggests that internal search is a tool that really doesn't help the user. Lots of people argued with him on this point, too.

Pogosticking is avoided by giving users enough information on the home page so that the user can accurately predict what is on the page. if you are talking about product pages, give basic specs on the gallery page so they can compare products without pogosticking.

He found that catalog pages that were limited in length (pagination) were less effective than a big, long scrolling page.

Listing is important. Alphabetical order is basically random order. People assume that things that are visually grouped together are somehow related in terms of content.

He's talking about how categories should be shared within an industry. Perhaps we are having trouble because our site serves a somewhat different purpose than other NPOs. We are creating a new set of categories, forcing users into more learning.

Most people find info by browsing the categories. Number of people who find things via features were minimal. This is not true, however, of news sites.

When he does usability tests, he starts by interviewing testee to find personal context and trigger words. He then uses those trigger words to perform the tests. No two tests are exactly the same.

:: Representing Many Voices: Challenges in Government Site Design

http://www.asist-events.org/IASummit2003/cs9.shtml

Social anthropologist who builds web sites for people.

Was tasked with getting 14 teams of arrogant hard-core academics to use a CMS. Talk was from the perspective of how to get the groups to contribute to the web site. The site hasn’t been launched yet, so she has not data about usability of the site they are creating.

She looked at her site as a community of communities (different, unique audiences). She took a “Geo-Cities” approach to site design.

The research teams were not cooperating, but she had to both convince them to work and present their research in a way that hides the differences.

She demonstrated that she could use the CMS to make changes on the spot. That overcame a lot of cultural barriers. Frequently she would go back and undo the changes but it allowed a better dialog. Less defensiveness.

Her system allows full control of their areas.

This isn't user-centered design. It's author-centered design.

Site will be live in September. She has no idea what will happen when she cuts people loose on this sucker.

Sunday, March 23, 2003

:: Persuasive architecture

http://www.asist-events.org/IASummit2003/persuasion.shtml

MAP-Minerva Architectural Process

http://www.grokdotcom.com/map.htm

Process that helps draw architecture from clients

Software projects have a 70% failure rate. This has been true since the early 70s.

straightforward tech project management

:: Kids in the midst

http://www.asist-events.org/IASummit2003/kids.shtml

Panel lead by Kathryn Summers

Difference between adult and child expectations for the web:

Adults - hard mastery

Kids - soft mastery (how you learn first language – see studies at MIT)

Adults - distinguished between online world and real world

Kids - no distinction

Kids did not really give good design input until they had the tools in front of them and were designing their own things. Before that, they just talked about standard adult design issues.

:: Experience levels in relation to usability

http://www.asist-events.org/IASummit2003/EEL.shtml

Study compared perceived and practical usability based on experience.

Found that expert users were faster, but made just as many mistakes.

Expert users were much more likely to overestimate their own effectiveness and NEVER used help. They were also more satisfied overall with the usability of the system.

Found that recommendations suggested by users aligned with Nielson/Norman standards

Problems were labeling and inconsistencies.

Experts and novices reported difficulties with the same tasks.

One interesting issue was the one where people filled in the wrong field because of labeling issues. Users were making mistakes and didn't realize they were making mistakes.

Used a program called Camtasia (http://www.techsmith.com/products/studio/default.asp ) that records screen captures, clicks, audio to an AVI.

When he presented this to the management team, they asked "well how often do people really do these tasks". He used log analysis to show how many times people actually those tasks that tested badly. This was very effective in the meeting.

Ultimately, the decision was made not to make changes because other business priorities took precedence in terms of money and staff effort.

:: Impact of behavioral-based research on design

http://www.asist-events.org/IASummit2003/impactBehavior.shtml

Ask users what their goals/expectations are about the site rather than just relying on business goals.

"Site traffic analysis gives you the most common and least useful information."

EasySort (EZ Calc) - freeware that helps with cluster analysis for cardsorting.

http://www-3.ibm.com/ibm/easy/eou_ext.nsf/Publish/1876

He uses card sorting to test top-down analysis. This shows where the analysis was wrong.

It is useful, but ambiguous. There's a lot of interpretation that you have to do.

Page rating system from OpinionLab -- probably too subtle for our purposes. Data visualization is pretty cool, though. $10K to $100K

http://www.opinionlab.com/

Click Link analysis by Keystone (service).

:: Arts and Crafts Movement
http://www.asist-events.org/IASummit2003/spirit.shtml

Michael Magoolaghan

Arts and Crafts was a way of life (moral standard) as well as a visual approach. Meant to be "medieval"

Ideals: Simplicity, honesty (it is what it looks like), utility (form should follow function, but it also needs to feed the soul), organic, craftsmanship (includes innovation within convention), harmony (a simplified and unified environment)

Used structural ornamentation - using joists and joints exposed. Wonders out loud how we could do that with web sites.

Treat compositions of a web site as building on a series of harmonies

Back to curiousLee

IA Summit 2003 Notes by Amy Lee

alee AT aarp.org

Saturday, March 22, 2003

Mark

Susan

Andrew

Sunday, March 23, 2003

^