Done Bright!

Just took A List Apart's 2008 survey for people who make web sites. As surveys go, it's not bad; but they should learn to size their text input fields to indicate the amount of text input expected. I mean, really — a 40-character field for age? :-)

So I uploaded Firefox 3 this morning, happily jumping on the bandwagon to make this the biggest one-day download in history (so far). I've liked Firefox 2 for years, and have set it as my default browser on all the systems I use (not only the Macs I use all the time, but the Windows OSs I run inside Parallels). I've played with v3 off and on all day, and I am disappointed.

Here are my biggest first impressions:

The tabs are too dark.

The tab on top is not much lighter (or should I say "less dark"?) than the other tabs, and they are all too dark to provide adequate legibility for anyone over 35. The default theme must have been designed by teenagers.

(Figure 1 illustrates this problem and the next two.)

The tabs hang from the bookmarks toolbar instead of being attached to the page content.

Tabs as a user interface design pattern were originally designed to mimic the tabs on file folders, where each one sticks up from its folder and is connected to it, so that the person could identify the desired content and bring it into view by selecting it. No longer is there a tight visual connection between a tab and the page it represents.

Now, you might argue that those nice folks at Mozilla deserve the benefit of the doubt, as you point out that the tab on top overlaps the bookmarks toolbar by a few pixels (approximately the distance between the bottom of the tab and the top of the web page) and speculate that it's a positioning error and that the tabs are actually supposed to touch the top of the page. However, should you do that I would urge you to look closely at the bottom of the tabs.

They're rounded. Yes, the bottom edges of the tabs have rounded corners.

Do you still argue that they were intended to be connected to the page? No, I didn't think so.

Simply put, Firefox 3 breaks the tab metaphor.

The new "back" button is flat weird.

I like the concept of having the entire history — both back and forward — available in one list, with a clear indication of which one is currently being viewed. But I don't like the Firefox 3 implementation of it. They've done some things right — made the "back" button larger than the "forward" button (remind me to post my elevator-button musing here sometime), grayed them out when there's nothing to go "back" or "forward" to — but the whole thing just looks peculiar, like something you'd slide rather than just click. And the dropdown arrow looks as if it's recessed, almost hiding behind the "forward" button.

The "back/forward" button complex just doesn't work, visually.

Some add-ons have stopped working.

This is to be expected in a period of transition. I trust that the ones that matter will be fixed.

Some of the "favicons" no longer show up in the tab areas or the address field dropdown.

Some do. Go figure.

Most of the themes are incompatible with the new version.

Some do work with it (perhaps they were developed specifically for it), but they can act pretty weird. I downloaded two of the new themes to try to fix the tab darkness problem, and both of those eliminated the scroll bars. Change a theme and poof! no more scroll bars. Now, I can always use that little rollerball on my mouse (note to Windows users: it's not a wheel)... but I don't like the rollerball thingie and I can read the tab labels if I squint (my eyes are better than most people's at my age)... so it's back to the default theme I went, cursed with dark tabs but blessed with scroll bars.

Who says you can have it all?

The new approach to auto-filling the address field is both good and bad.

I approve of the attempt to give users more information in the dropdown list than just the URL, but sometimes they overdo it and sometimes it's actually incorrect. In particular, the description that appears above the URL doesn't always match what will come up if you click that item. For example, in the list shown in Figure 2, the Google Maps link says it will take me to "Vicolo del Bologna" in Rome, although in fact the link goes to Google Maps Italy without the detailed data. Evidently, Firefox is just showing the contents of that window's title bar, regardless of how well it matches the URL.

Firefox 3 does seem to be faster than v2.

But not all that much. Certainly not enough to outweigh all of these usability problems.

The verdict: I'm going to stick with Firefox 2 on my laptop — at least until v3 matures a bit — and may even retreat to v2 on my desktop.


Figure 1. Top of Firefox 3 window, showing tabs, back/forward button complex, and address bar.


Figure 2. Address bar dropdown list

The American Library Association has mentioned our work in a press release describing its plan to unveil its new web site at its 2008 Annual Conference this June (link will open in a new window). I've been the technical lead on this project for almost two years (ALA has taken the time to do a thorough job!), first as a UserWorks employee and then (since June 2007) as a subcontractor to UserWorks. The description of my role in this process is in the Luminanze portfolio.

Labels: ALA, American Library Association

The annual MatriarCHI lunch was held on Tuesday at da Sergio Trattoria in Florence. I have my friend Fiona to thank for the reference to this restaurant, as she and her husband Enrico are members of the Slow Food society and she read me the list of the Osterie d'Italia in Florence. The lunch was a wonderful meal shared by a congenial group of more than 15 women. The trattoria proved a little hard to find, as we had to peek between the stalls of the outdoor market in Piazza San Lorenzo, but we managed, and a good time was had by all.

From the MatriarCHI description on Yahoo Groups: "The members of this group are all the significant women who have contributed to the nurturing and development of the field of human-computer interaction. They have served in leadership roles either by volunteering their time and energy to the workshops, conferences and societies that foster the growth of this area or to the creation of theories, models, systems and significant scientific accomplishments in HCI."

Labels: CHI 2008, MatriarCHI

Last week, the European Space Agency's Columbus Attached Laboratory finally made it into space. I worked on this project some 20 years ago but hadn't been keeping close track of it, so it was an especially nice surprise to hear of its successful launch.

What was my role? you may ask. Human factors, of course. I supported AERITALIA (now part of Thales Alenia Space), the company that built the lab module, in the human-centered requirements and process for the software and hardware of the Crew Work Station. Although I have no idea how much these things evolved after the end of my participation, I feel proud to have contributed to the effort.

In bocca al lupo, Columbus!

One of the most interesting (and most crowded) sessions at the Interaction 08 conference was Bill DeRouchey's The Language of Interaction. One of the bad designs he showed was a Farecard machine from the Washington, DC, Metro system (my local subway system; see slide 22 at the above link), which he captioned "Inspiration from what intimidates us". (No kidding!) I enjoyed the reference to something from home.

"We are interaction designers during a time of rapid technological change," DeRouchey says, "placing us in the incredible position of collectively creating and curating a new language, the language of interaction."

But is it really new? Certainly the concept of an interaction language is not new. For example, cybernetician Gordon Pask published his theory of conversation more than thirty years ago, and he applied it to communication not only among people but also between people and machines. (See Paul Pangaro's explanation.) And the first course I took on interaction design — way back in in 1984 — addressed it in terms of designing "human-computer conversations". Conversations, I submit, can be accomplished only through a mutually understood language of some sort. So no, thinking of interaction as involving a language is not new.

However, language does evolve, and here's where I think DeRouchey's point is most relevant. Each new technology — each new interaction medium — adds to this language. Some even subtract from it, as they supplant older technologies.

"We are the custodians of this language," DeRouchey says, "creating and curating it organically. We need to start seeing it everywhere and learn from each other as we define the future relationship between people and technology."

Amen, Brother Bill.

So I'm here at the Interaction '08 conference in Savannah. I'm learning some new things, being reminded of some things I already knew, and meeting lots of people, some of whom I have known electronically for some time. I've seen people blogging this thing (especially David Armano, who sat next to me at breakfast and who showed me how to use my Treo to Twitter), so I won't go into detail until I've had more time to think about it, but I wanted to share some random thoughts and ideas that have been having a mashup in my head for the last couple of days.

Yesterday afternoon's three-hour workshop on Agile and Interaction Design gave me a couple of ideas I think I can take back to a current project, mainly to help turn an ad-hoc "agile" process into a more intentional one. I had had little sleep the night before and the afternoon is usually my low-energy time, but the combination of engaging presentation style and group exercises kept my attention level reasonably high and I was glad I attended.

At last night's reception, I spoke at some length with Darja Isaksson, who, when she found out how long I've been in the field (more than 25 years), asked me what it was like when no one was doing it and we had to fight for human-centered design. I commented briefly, and then she asked an interesting question: What did I see, she wanted to know, as the advantages and disadvantages of having so many people in the field now and so much attention paid to it. "I don't see any disadvantages," I started out, but then I changed my mind. The disadvantage, I told her, is that the field is so full of people who don't have any background in the human factors research that started it all, and they can't explain why something is a good design. (For example, they may know that a lot of movement on the screen is annoying to users, but they may not know that peripheral vision is more sensitive to movement than is foveal vision — hell, they may not even know the word foveal — so they cannot explain why it is distracting.) Finally, Darja said something in passing that gave me an idea for a talk. (I'll say no more on that until I have it fleshed out somewhat.)

Today's most useful session was Dan Brown's on Concept Models. His "concept" (pun intended) was eerily reminiscent of the Conceptual Design activity I learned in Jim Foley's class at GWU back in 1984, although Brown's approach is more fully fleshed out, and I found his different styles of diagrams helpful. I'm going to start using that one right away.

Oh, and I forgot to say — I'm surrounded by Macs! <grin>