back to tim's mas.123 page
For the image maps assignment, we set off for Harvard Square. We spent about an hour walking around, taking pictures of areas we thought were interesting. This included a number of spaces that were particularly people-friendly-several places clearly intended to be gathering spaces, a park-like area, and some wide, and leisurely walkways that invited pedestrians to stroll along the storefronts. Seeing everyone out walking along the street, I noted that Harvard Square is a very pedestrian-friendly environment-but, having tried to navigate it in a car a few times, it's certainly tough to get around in a car. One-way streets are connected almost at random, and roads simply don't go where you'd expect-such as towards MIT on Mass Ave. We took a lot of pictures, trying to thoroughly document some of the spaces around the square from different approaches and so forth. We also went through a correspondingly huge amount of batteries-three sets of four AA's. After some initial trouble, our only problem with the camera was that it would get locked up every third picture or so, requiring us to remove the batteries and reinsert them. We found the "video viewfinder" useless (we looked through the real viewfinder instead)… so if the scripting language allows it, that feature could be safely removed.
After Erik imported the photos, we sat down with the Image Map software to
look at our photos. I 
was
disappointed that only a small subset of the (admittedly huge number of) photos
that we took could be viewed in the software. In particular, some of the spaces
that we had documented around Harvard Square were totally omitted from the set
chosen. The set of pictures chosen focused on the area around the T station
entrance, thus guiding our investigation towards that area. It is interesting
to note that our walk through Harvard Square specifically looked at a variety
of areas around that central space, yet the choice of pictures to examine further
naturally guided our investigation.
As we began to explore the historical pictures which the software presented to us, we found that the first striking feature which had changed over time was the T station entrance itself. At first we couldn't find the entrance at all in pictures before 1980, but eventually we came to realized that the structure which is now Out of Town News had actually been the T station. We were especially excited when we found still older pictures indicating that there had been an older, round subway station at the same location. We were at first uncertain whether this round structure was a subway station at all, until we searched on the internet for "MBTA History" and found that there was indeed a subway station at Harvard from 1912-1981, at which point it was replaced by the current station. We even found a great picture from 1980 showing the underground construction of the new T station.
In this exploratory phase, we simply switched between pictures, not grabbing
pieces of them to place
into models, as we had not yet decided on what we wanted to pursue in more detail.
Thus, we found that it was very difficult to go back to a picture that we had
seen previously. It seemed we wanted a "back" button to go back to pictures
we had seen previously, but since we were (somewhat) systematically looking
through all the pictures that the software was giving us, a normal web browser-style
"back" button or history list wouldn't make much sense. Instead, perhaps it
would be useful to be able to construct a list of photos that you find interesting
and be able to select them later in your exploration. This would make it easy
to, say, switch between two historical photos which were accessed from different
contemporary photos-a process which is difficult in the current software.
As we continued to explore, we moved beyond the obvious change in the T station and started looking at changes in the use of the space. In numerous old pictures of Harvard Square, we saw a much more vehicle-centered space. One particularly striking image showed the T station (which is now Out of Town News) surrounded by a circle of sidewalk with roads all around, clearly making it difficult and dangerous for pedestrians to reach the station. Other images showed pedestrians crossing the road surrounded by cars, trolleys, and trucks. We began to look for changes in things such as crosswalks and sidewalks, noticing that crosswalks only seemed to appear in the 1970's, and that they were recently changed to the brick crosswalks in the square today-this seems to have occurred with the 1981 renovation of the T station. We also noted fewer buildings in earlier photos with stores which seemed to stretch out onto the street. The general picture seemed to be that since 1980, the sidewalks of Harvard Square have become much more pedestrian-centered, providing public, social spaces that seemed to be unavailable in earlier periods.
The feature of the Image Map activity that seemed most effective was the core of the project-that we chose the area of investigation. Despite the guidance that Erik's selection of our photographs provided, the fact that we had actually walked around Harvard Square to take the photographs made the experience of using the Image Map software much more powerful. Even though I'd walked around Harvard Square before, I still had a different experience as we walked around with the camera because we had been instructed to look to the spaces and consider how they work architecturally and socially, and consider different approaches to a space and look for signs as to how they had evolved over time. This examination of the space in its present state was very helpful as I looked at historical photographs, allowing me to form clearer comparisons between the photographs and the understanding of the space that I developed while taking the pictures.
I liked the idea of using the software to pull together similar parts of different
photographs into a "model," though I found annotating that model to be difficult
because of the small, fixed size of the text window and the 
inability
to make general comments on the page itself. I was also somewhat disappointed
that the only relation allowed between elements of the model was a directional
arrow. This seemed to clearly lead towards a chronological view, which seems
quite limited. I would also like to be able to have grouped pictures by drawing
a shape around them, annotating that shape, drawing arbitrary arrows (say between
groups) and annotating the arrows. This is only a particular suggestion-I think
that in general the tool would benefit from less constrained means of relating
images in the model. We also had some trouble with the method of selection used
in the panel used to compare entire photos. The trouble arose because of the
two selections models used on the screen. My suggestion was to change the method
of selecting the contemporary image to match that of the historical images,
and the put the selected contemporary image in the upper-right-hand corner,
making it the last in the historical timeline (and the historical images should
be correspondingly put in chronological order.)
I liked the idea of setting up a hierarchy of annotations to use in the investigation.
However, we were unable to make use of this feature because of our caution after
the program froze a couple of times as we used such a complex hierarchy. As
we tried to use a wide hierarchy of annotations, I found it difficult to pursue
what seems to be the natural strategy-find a picture which contains interesting
features about the things you're investigating, select the interesting features
of the photo, and place them into the appropriate models that you're constructing.
However, because you can only add your selection to the currently open model,
this requires a great deal of switching between tabs to add all 
the
selections from one photo. It would seem useful to be able to change the currently
open model from the floating "info" window. After having the fear of God put
into us by a couple of program crashes in which we lost the annotations we were
constructing, we made only limited use of the flexibility afforded by this hierarchy-we
basically constructed one model by searching through the photographs for interesting
features, then took a screen shot of the model and started on another model.
Overall I enjoyed the image map activity. Though the software needs some further interface work, actually going out and taking the pictures gave me a strong personal connection to the analysis that we later did with the software. The notion of letting the user choose the object of investigation-both in the historical photographs that would be examined and the features of the photographs which the user chooses to explore-worked very well. The activity was a clear win over the more traditional method of simply dropping a bunch of old photos on a tabletop and looking at them in both the personal nature of our explorations and the way that the software forced us to be clear about the nature of the conclusions which we were drawing.
for more information about the Image Map, see:
finally, back to tim's mas.123 page