Some of the readership of this blog are people who work in the humanities — literature, criticism, art, museology — and some work in technology. Some work at the intersection of both, like me. So I figure this is a great place to pose a question that hit me like a hammer today.
Are there problems in the humanities that can only be solved by a supercomputer or some sort of distributed massive computing platform?
Anything that requires heavy doses of processor-crunching? Large corpus text analysis or image analysis? Help me here.
Protein folding, deep space radio astronomy, thermonuclear explosion modelling, meteorological forecasting and brute-force decryption cannot possibly be the only uses for supercomputing.
Do tell, do tell!
i’m not sure if it quite fits your description, one example of an excellent emerging application is happening at my school: wordhoard and its accompanying calculator. the goal is to automate typical philological queries to aid in close reading of medieval and early modern texts.
Matt, thanks. Very cool project. That is the type of thing I am after, though I am thinking of tasks that can only fruitfully be performed by computing capacity that exceeds a typical PC.
I should say that two things — this article, which talks about supercomputing in the service of textile preservation and this project, which facilitates grid computing in the life sciences — have caused me to ask.
There are a giant bag of humanities related questions you could throw some cpu’s at, but would also need a little programming too. One problems that’s on my to do list: no one knows the exact order of Plato’s dialogues (at least last time I checked with those who track this stuff). If you could grab all the translations, “originals”, and so on you could look for patterns of language use (be it from the transcribers or whomever) or perhaps develop some interesting thematic or concept maps and then crunch on them for a while to get some probabilities. Scale this up for other ancient sets of work, especially those with more data to analyze.
Oh, another one. Take all the versions of a classic work such as Ulysses, which has many proofs, edits, etc. and see if you can trace the evolution of ideas, writing styles, ingenuity and so forth. Insight into genius?
Okay, now let’s go raster – look at the progression of say, ALL of the French impressionists and look for commonalities among their approaches, colors, subjects, object sizes, etc. in the works themselves. Added bonus problem, map all the paintings on a map from their real world inpirations (Giverny…) and of course provide ways to slice that up over time. Influences, seasons, popular locations (e.g. were most urban impressionist painting subjects within X feet/miles/yards of the Paris Metro stops at the time of creation?). And of course, multiply this analysis by adding in all the sketches, drawings and non-final versions of a work too.
Anyone want to give me a grant?
Don, what a great list. You’ve clearly thought about this at length. Do you know of any examples of projects that are actually ongoing? If so, know of any that are being tackled in a distributed or grid fashion?
Text-Grid is an example:
see their work on “interweaving” historical dictionaries with texts–an example of combining large datasets.
Also, IMIRSEL: http://www.music-ir.org/evaluation/m2k/index.html
I’d like, though, to find more examples of the humanities research questions that could drive this initiative. It is interesting that the Unicorn tapestry example, while a challenge to the mathematicians, does not “solve” the _humanities_ question, at least in the article mentioned here. The tapestries are still an enigma.