What is/was MESL?
The Museum Education Site Licensing project is/was a project set up to explore the use of Museum owned digital images in an academic context. This was accomplished through big bureaucratic manuvering (necessarily so because of the complexity and scale of the orginizations involved) that resulted in the distribution of identical sets of big data-chunks including about 10,000 digitized images of holdings from 7 museums and variously organized image-specific associated text to 7 different academic institutions for subsequent deployment to members of those institutions through some sort of web-based format.
What so interesting about that?
Lots of things. There are many issues addressed by the project, ranging from fair use legality to database design, and most importantly, the project marks an important step at subverting pure commericial intrest by establishing a model for the flow of cultural history between and among institutions ostensibly focused on goals quite outside the market economy.
But relative success at that sort of goal is a difficult thing to measure, especially in real time, so instead we'll focus on something more immediately obtainable: relative usefulness/usability of the material as mounted by each institution.
Why my MESL page lacks images:
No good reason. I've struggled with converting a little over a dozen Mac Pict files (mostly of the atrocious results I collected during my "Violence search") using every piece of software available in the SIMS lab, and I just can't turn them into jpegs, gifs, or anything else useful. In fact, everytime I attempt to open them using Photoshop (or anything else) I either get error messages or fuzzy static-like patterns. They're still sitting on SIMS server, in this very directory, should anyone decide to play with them. Instead you'll have to rely on your imagination and my descriptive powers.
My Approach to MESLing:
I wanted to do some basic level searches of the sort that a student using MESL for the first time might perform. I selected 3 search topics on the following bases:
The VIOLENCE search -
Early on I started doing searches based on what I arbitrarily decided might be "loaded words". I was interested initially to see whether data base managers would choose to deal with material associated with culturally contested concepts such as sex, violence, abortion, gay rights, etc. This tack was not very productive for reasons I delve into below, but I kept the seach for violence anyway. The reason was this - I found the same Roy Lichtenstein print on many of the searches. Interestingly, though all the images on all the sites were derived from the same set of original digital images, some looked horrible mangled and misfigured. I decided to keep the search as an example of an issue not directly related to database architecture - the issue of content quality control. Across databases the same one or two images were consistently selected.
The HELMET search -
This, I thought, would be my value-free simple search across all the databases to show the actual continuity and similarity of results. in fact, some searches produced as many as 15 matches, and others produced 0. Most sites produced between 11 and 15 matches, and the 11 were constant across results. The highest return rate, 15, was achieved at the University of Illinois site which searched across "ALL" fields of information including "Artist/Creator", "Title", "Nationality/Culture", "Object Type", and "Subject". Most of my other searches were intentionally limited to the "Object type" or "type" field, with the curious exception of Maryland. Maryland University produced the 0 matches result which astonished me - like the University of Illinois, only one search option was available to the novice user (i.e. one who has little or no familiarity with SQL). Instead of the University of Illinois explicit statement of what fields and infomation one was searching through Maryland offered only the somewhat opaque "Keyword Search" which resulted in a 0 match result so I can only assume that "Keyword" does not compare with any object type field. This renders the SQL-illiterate user unable to make any search based on the the type of object that the digital image has been made from.
The ANONYMOUS search
At this point I was interested in examining the ways people chose to organize and represent information associated with a piece of art. Many conventions seem to exist in different contexts for crediting a work done by an unnamed artist depending on the circumstance. I wondered to what degree any consistancy was imposed on the collections.
Not too suprisingly, there was none. A particularly large mass of 18th and 19th century drawings had been credited to Anonymous and these cropped up regularly - most of my search results ranged in the high seventies and low eighties with a fairly consistent base of the same images. I tried to search in whatever category corresponded roughly to "Creator" or "Maker" or "Artist Name" at the various sites, but I was again prevented from doing so at U of Illinois and Maryland for search-design issues raised above. The Illinois search yielded 78 matches (on an ALL field search) and the Maryland Keyword search resulted in only 1 match found. The best theory I have been able to construct regarding the striking disparity in results is that the Maryland search engine, though provided initially with the same data fields from the various museums, chose to not retain whatever field usually held the information that the work was created by an anonymous artist. American University also had rather anemic results - only 5 matches found. I can only assume that they too failed to retain the appropriate field information.
Search data collected by site: