Seminar planning--general discussion area

From PublicWiki
Jump to: navigation, search

Caroline 11/3/04:

  • Brainstorms for structure of the seminar:
    • Possible formula: Start with a CS issue that is well-known to computer scientists but not to folks outside CS and then consider this issue's relationship to social goals, bringing policy, law, business, economics to bear on studying the relationship of the CS issue to larger social goals or circumstances (vague sounding, need better language to describe this). The idea is that a lot of the most intriguing interdisciplinary work in tech/policy/law/econ has come from simple observations about computer science that make computer scientists think "of course that's true," but are news to people outside CS. For example, Larry Lessig's latest brainstorm is his observation that digital technologies work by making a copy of everything (the "no duh" observation to computer scientists) and that therefore, copyright law perhaps should no longer work as it currently does by regulating every copy of anything (the consideration of the CS issue's relationship to larger social goals). Or Dave Dill, CS at Stanford, has publicized the notion that software quality is terrible and that software programs are not to be trusted. He publicized this notion as part of his crusade to have voting precincts not depend solely on DREs that do not produce a paper trail.
  • General thoughts on the topics for the seminar:
    • Should we study one technical issue and its connection to social goals in depth (i.e., improving software security) or focus on 3 or more topics? Advantages to one topic: more than a superficial understanding of the issue, possibly less work for organizers if we can become experts in just one area and if time crunch for brainstorming and developing other topics is an issue for Ben and Keunwoo, possible growth of the 590 into a larger project. Disadvantages to one topic: attract a limited audience, perhaps people will get bored of the same issue all term unless they care a lot about that issue, don't get to see what kinds of topics UW community is interested in. Advantages/disadvantages to discussing more than one issue are the reverse of these.
    • How do we find more of these topics?
    • Possibility of pursuing funding to bring in guest speakers, could be discussed in next meeting if relevant. This could be done as part of the 590 or separate but either way, I'd like your thoughts. As one of the many darts I'm throwing out there trying to see what we can do interdisciplinarily on this stuff at UW, I'd like to try to get funding to start up IT/public policy talks to continue what Ed L's done in his class. The goal would be to follow the above-described formula for interdisciplinary work, cover topics he didn't get to cover and open up this discussion to the many non-CS-PMP folks who would have liked to have taken his class. What would be great would be to bring in a computer scientist to explain a CS issue and a law, policy, or biz/econ expert to explain the implications of the CS issue in a tandem talk. So, for example, you could ask a CS person to talk about CS's ability to trace packets and then a lawyer to consider some specific privacy issues this raises. Or David Notkin could talk about the problems with creating secure software and an economist could chime in on whether creating an underwriter's lab that measured security, and thus conveyed information to consumers on how secure their software is, would produce consumer demand for secure software. Assuming objective replicable tests the underwriter's lab could use for this could be devised, something David could comment on.
  • Topics for the 590:
    • Ben's suggested topic from Oct. meeting: vendor liability and software quality
    • Keunwoo's: disk capacity and storage of intellectual property
    • Caroline's: reverse engineering law and CS: the professor in law I spoke with about this hasn't really begun his research so we don't have much to go on if we want to pursue this. There may also have been some uncertainty about what the CS issue to discuss here might be anyway?