Monthly Archive for November, 2005

To all the people on Planet Informatics, I re-did the jar template (which I had lost) and created head-in-a-jar icons for a bunch of the new people. If you don’t have a head or don’t like your head, feel free to download the photoshop template and make your own and then email me a .PNG file. Or email me a photograph and I’ll make a head-in-a-jar for you.

1. I just realized: Despite my beliefs hitherto, I have never eaten a peanut–only peanut butter.
2. I would very much like to see the new Pride and Prejudice movie. This has nothing to do with Keira Knightly.

What was the last thing you did with a computer that really mattered to you? I recently played a game of Go with my brother, connected with him over a thousand miles. That meant something to me. I have some photos I took at Brown County State Park on Saturday and I promised someone I’d send them to her but I haven’t gotten to it yet. That means something to me.

What about you? What have you done in this medium that you really cared about? What do you really care about that you are going to do?

Now look at the computer screen. How much of the stuff on the screen has to do with those things?

In my case, the answer is: nothing. There are dozens of words on my screen: File, Edit, View… Applications, Places, System. There are icons of globes with foxes on them, and little orange men. There is a “File Browser” and “Network Settings”. Even now, as I am typing this, it’s not my weblog entry about how computers suck, it’s “Unsaved Document 1″.

When you open up your journal to write, what are you greeted with? Maybe a blank page. Maybe yesterday’s entry. When you pick up an old school camera, you see a viewfinder. It pulls you in and there’s your photo. When you walk into your kitchen you see vegetables and spices and pots and pans. But when you walk into computerville what are you greeted with?

Utter garbage.

How’s this for an idea: how about when you turn on your computer you are greeted with photographs and weblogs and letters from your family and notes from your friends. When was the last time you looked at your old digital photos? Or read an old note from someone who cared about you? When was the last time you stumbled across an old friend on your computer?

It happens, but it’s rare. All of these things are hidden behind menus and icons and bookmarks and hyperlinks. From behind the keyboard you can find anything, but things almost never find you.

Not only that, but when you do decide you want to go find an old friend or send someone a photo, you have to slog through menus and files and saving and opening and uploading. You can’t just bump into someone in the digital hallway and hand them a photograph. You need to massage the digital bureaucracy into doing it for you.

I’d like to change that.

How? By starting over and putting the things that matter down first. It reminds me a little bit of Marty’s “big rocks” analogy. He says our lives are empty vessels that we fill with stones. We put in little rocks–things like assignments and errands and favors that are important, but perhaps not the most important things in our lives. We put these little rocks in our vessels and when it comes time to put in the big rocks: our passions, our loved ones, our sources of pride, we have no room left for them.

We need to put the big rocks in first, is the moral. The things that matter most go in first and the little rocks find room where they can.

That’s how we should do design: figure out what really matters to people and make it work first. Everything else is secondary.

We usually start designing web sites by creating the navigation bar. Who cares about a navigation bar?

Food for thought.

In genomics, brute force searches are often infeasible. They find either no matches or too many. So bioinformaticists need to use clever heuristics to narrow the search space or the results. This is an immensely creative process that requires intimate knowledge of both biology and code. This is what makes the work bioinformatics do with biologists so collaborative.

Unfortunately, such collaborations are few and far between. There is an shortage of bioinformaticists good enough to architect novel search methods. And there is a lot of work for the bioinformaticist to do even without getting creative with the search algorithms. The bioinformaticists has to wrangle with UNIX and the large volumes of data. Once that task is done, they often just accept the defaults of the search application. They may not even know what parameters are available to them.

After they get results, they may selectively learn what parameters of the search algorithm can be tweaked, working with the bioinformaticist to figure out where it might be important to do so. But there is a disconnnect here: the bioinformaticist understands the computer system and (partially) the software and has the job of translating the results into a form the biologist can understand. The biologist understands the physical biological model and has the job of translating their knowledge of biology into directives for the bioinformaticist. As a result, it is hard for either party to really make insights that would move the research course out of the mainstream.

One possible intervention would be to create a visualization of how the algorithms work. This would be instructive for both the biologist and the bioinformaticist who is not intimitely familiar with the code. It may enable the biologist to find creative ways to change how the code works. And even if the biologist can’t see exactly how to change the code, they will be able to use the visualization to help them communicate to the bioinformaticist what they need to do and the bioinformaticist can figure out how to do it.

In general, we think software should be opaque: hiding from the user what is going on under the hood. But in this case, it may be that more transparency is needed. This may be true in general–that increased transparency is a key method for enabling creativity.

Erik (the new head of the HCI program) just sent out an announcement of a new journal he is editing, which “strives to promote transdisciplinary design research.”

Transdisciplinary? I am thinking, I have heard of interdisciplinary, but transdisciplinary?

Thankfully, Google is always there to answer our questions, and it found a useful mailing list post which cleared up the confusion. It seems that interdisciplinary research is about combining ideas from multiple fields towards a single pursuit. Transdisciplinary research, by contrast, actually connects multiple disciplines together. It is about developing new ideas that relate previously unrelated disciplines under one roof.

So, I would say that human-computer interaction tends to be interdisciplinary: we draw on psychology, computer science, anthropology and design, but you generally don’t use HCI to draw connections between, say, computer science and anthropology (though it would be possible to).

Cognitive Science seems to be more transdisciplinary. It is focused on uniting things like artificial intelligence, psychology, computer science, and robotics under one umbrella. In this context cognitive science, computer science and psychology actually become branches of a single pursuit.

I also want to mention that I recently crossed the 500 entry mark on this blog. Incredible aggregate effort? Monumental waste of time? You decide!