12 Months of HCI

A notebook for Carnegie Mellon’s Masters of Human Computer Interaction program

November 17th, 2007

Sony Factory

We went on a Human Factors field trip on Thursday to a Sony factory an hour outside of Pittsburgh. It was the most interesting part of the class so far. However, it was ironically interesting more as a lesson in global economics and the manufacturing business than in Human Factors.

I’m sure it must have been impressive a few years ago when thousands of people worked there manufacturing big, heavy CRT TVs. It made sense then because TVs were so heavy that it was economical to make them in the US and save on shipping. But now that flat screen TVs are so popular, its cheaper to make TVs in Mexico or China and ship them to the US because they are much lighter. So now there are only 2 production lines at this factory. And robots are increasingly replacing the workers that remain. In fact, one had to wonder to what extent the Human Factors issues that we studied haven’t driven up the cost of US manufacturing and contributed to those jobs going to countries unconcerned with cumulative trauma disorders and people bending over to pick up heavy objects.

November 15th, 2007

The Affect of Analog Clocks

I admit that I’ve never been very good at reading analog clocks. I’ve always used digital ones myself. But when I moved to Pittsburgh and no longer had a cable box with a digital clock on it to tell the time in the living room, I discovered that digital wall clocks are really, really hard to find. So I got an analog clock for the living room and between that and the Brain Age clock training game, me and analog clocks were slowly mending our relationship.

Then the subject of displays came up in Human Factors and the teacher insisted in class that analog clocks were faster for telling the approximate time than digital clocks were. Alot of people in class disagreed with her and I really disagreed with her, verbally, even after she gave further reasons for analog clocks being faster for approximate times. Granted, we both agree that digital clocks are faster for telling the exact time. And I think that part of the issue is that my brain doesn’t work with approximate times. It wants to know the exact time and will sit and look at the analog clock to calculate it.

Anyways, the interesting/unfortunately part of all this is that I’ve noticed my time telling skills on analog clocks have gotten alot worse since then. When I look at my clock on the wall, all I can think about is how many extra milliseconds of cognitive processing time it takes to compare the two hands of the clock to figure out which is the minute hand and which is the hour hand, the time needed to realize that the hour hand is slightly before the”10″ and that its thus only 9:55, not 10:55, etc. Its as if all my gripes with the class have been transferred to the clock and my time telling ability has suffered for it because my mind focuses on the argument against analog clocks rather than figuring out how much time I have before class.

October 11th, 2007

Heuristic Evaluations

Its a fancy sounding term. But when it comes down to it, Heuristic Evaluations are basically going through an interface and writing down all things that match a list of abstract good or bad features. Its a bit more formal and structured than that and it helps if you are an experienced professional. But thats the basic idea.

Its easy to see how this simple idea can be a powerful and cheap way to discover usability problems. In some ways its just a more rigorous and formal version of the way I would find usability issues as a developer. But it reminds me of the usability report I once saw that the Nielsen Norman Group did for a site I was working on at the time: Merck.com. Although I can’t find my copy now, I remember at the time thinking that many of the suggestions in the report sounded reasonable in the abstract, but didn’t really seem that relevant or applicable in this specific case. Apparently I wasn’t the only one who felt this way, as I don’t remembering talking about it that much at work and I’m not sure if any recommendations of the report were actually implemented.

Using this Heuristic Evaluation for the first time myself, I can see myself finding “issues” that I can justify with the given list of usability principles and thus quickly record. For example on the site I’m evaluating, having all these ads clearly violate the “aesthetic and minimalist design” principle (even though they are Google text ads). However, I would never think that this was an issue normally and I’m not convinced that its really important. After all, the site has to make money or at least defer some of the server costs and the ads are not that distracting. I’m only recording it because I’m explicitly looking for things off of a given list and the goal of the game is to find as many as possible, even if they are minor infractions and don’t really cause any harm in the given context. I feel that I’m thinking like the usability experts who wrote that Merck.com report that I scoffed at and ignored so many years ago.

This seems to be a somewhat known issue with Heuristic Evaluation since it was discussed in class that it tends to find many issues that don’t have a noticeable impact on usability. However, my criticism is not so much that it tends to find minor issues. I’m all for more polish in interfaces. It seems that it manufactures artificial issues that are only of concern to people temporarily confined to reading off of a list of ten usability principles.

September 1st, 2007


For 5 years I worked with a group of mostly Java developers who had to drive about an over an hour each way once a week to the client’s office. In this time I managed to avoid ever getting behind the wheel of a car (I’m phobic of driving in big cities, especially New York). I also mostly managed to avoid touching Java. This had less to do with a lack of Java knowledge (though it is not my strength), and more with a dislike of the type of work that Java lends itself to. When I did make the occasional commit to Java code (usually a minor change, I only remember writing a new class once), it often elicited some sort of joke from a co-worker (or myself) that I had actually written in Java. So when I sat through the first part of an ad-hoc “refresher” session on Java for my Software Architecture class, it was reassuring that I didn’t really learn anything new.

Read the rest of this entry »

August 31st, 2007

Elective Confusion

Yesterday, just when I was starting to feel more stabled and settled into Pittsburgh and the HCI program, a little bit more chaos entered into my day.  I went to the first class for my highly anticipated “Evaluation of Computational Cognitive Architectures”class (what an impressive sounding name!).  Right after I sat down I found out from a classmate that it wasn’t really a class and even though it had already been approved by my adviser I couldn’t get credit for it.  I wanted to take some classes in Cognitive Modeling and this seemed like my best (and only) option to do so this semester.  I knew that it was going to be a seminar style class.  I assumed that we would have readings every week that we would discuss in class and then do a significant paper or something at the end.  This is how a previous seminar class in undergrad was structured.  It turns out its really more of a series of presentations by various speakers.  You can ask them questions as they go, but its not really a group and discussion.  And its definitely geared towards professors and PhD students.  And most importantly there is no work to do.  Thus, my previous approval for this elective was revoked and a mad scramble to find a replacement is underway….

Read the rest of this entry »

August 27th, 2007

First Day

I’ve just finished my first day of classes at CMU and so far so good.  During the orientation the previous HCI graduate students said that on the first day of the HCI Methods class last year the professor told everyone that they were already a week behind in their reading.  Thankfully that didn’t happen today. In fact we have no homework (yet).  The class still looks like alot of work, but hopefully it won’t be as insane as it was made out to be.

My biggest issues have been with the non-CMU parts of Pittsburgh: Verizon screwing up my internet connection, my landlord loosing track of payments I sent them (twice),  Pennsylvania alcohol laws that seemed designed to make the purchase of alcohol an all around unpleasant experience.  But the CMU parts are all good.  My fellow MHCI students are smart, friendly, and diverse.  The classes look interesting and useful.  The trickiest part of the semester looks to be the logistics of scheduling meetings and work for the various group simultaneous group projects.