The Quality of Information, Infosys 290 Section 10 Paul Duguid and Geoffrey Nunberg
Class Sequence

------------
Paul Duguid

Geoffrey Nunberg

GUESTS
John Lamping

Peter Lyman

Clifford Lynch

Steve Weber
------------

Class blog

Personal library
Previous week

September 29: Rotten Information (2):
Pornography & Racism


For Geoff's class slides go to the page for the 10/6 class

Primary Reading

Rajagopal, Indhu & Nis Bojin. 2002. Digital Representation: Racism on the World Wide Web. First Monday 7.10. October.

Secondary Reading

Nordland, Rod, et al. 2001. The Web's Dark Secret. Newsweek. March 19.

Paul, Pamela. 2004. The Porn Factor. Time. January 19.

Rogers, Michael. 2003. Practical Futurist: Are Computers Wrecking Schools. Newsweek. Oct 14.

Search Assignment


In the second half of this class we will discuss the following search assignment: Imagine that you're a typical UC Berkeley sophomore and have just been assigned a paper on some plausible topic over which there is some division of opinion-- whether the relation of Vitamin C to cancer, simplification of the tax code, electronic voting, heterosexual spread of AIDS, Ebonics/Black EnglishÉ. pick whatever you like, so long as it's the kind of topic where people recognize the relevance of "authoritative" opinion, even if they disagree as to what it is or who is entitled to hold it. (That is, don't pick the question of whether the intentional walk rule should be changed or whether hip-hop is a force for Good or Evil)

Like most of your classmates, the first thing you do is go to Google and enter a search query or two. You come up with a bunch of pages of varying degree of authority or reliability. Then be prepared answer the following questions:

1. What makes for the impression of authoritativeness in a document found on the Web? Give specific examples of institutional cues, domain names, graphical and typographic cues, linguistic cues (right down to punctuation), cues in html code, and anything else that seems relevant.

2. How can the impression of authority be undermined? That is, what kinds of (actually) authoritative pages are likely to appear non-authoritative and vice-versa?

3. Which queries work best for finding authoritative information? Are there any general principles here? Does it help to look at pages that come up later, or to include non-relevant terms like "paint" or "Cleveland" that will effectively reduce the set to a more random sample? What happens if you include words like "authoritative," "official," or "debunk"?

4. How could the results be improved using multiple search engines? (For these purposes, one other will suffice.) E.g., what can you do using the proximity feature of AltaVista or some other feature that Google doesn't have?

5. Anything else we should be thinking about here?

We'll ask you to produce a short write-up of your results (which will be ticked off but not graded). Feel free to work in groups of 2 or 3.

Next week