On Thursday, August 11, 2011, Jonathan Rochkind <rochkind_at_jhu.edu> wrote:
>> Has the public wanted catalogs or bibliographies? From my research, and
>> also from my experience of what patrons want, including myself, I would
>> hazard a guess that in the vast majority of cases, they want
>> bibliographies because very few people who are interested in the text
>> and criticism of Huckleberry Finn want to see records for completely
>> different topics, e.g. Arabic calligraphy or tomes on the war of the
>> Spanish Succession. In a catalog, you will see these records that are
>> irrelevant to your needs, especially in a keyword-type environment,
>> while in a "Bibliography of the texts and criticism of Twain's
>> Huckleberry Finn" you will not. In catalogs, it is almost inevitable
>
> But this historical idea of a bibliography hand-crafted by an expert on a
particular topic -- it is highly unrealistic to think that such a
bibliography can be created for every possible topic a user may be
interested in.
>
Well, this statement is true if you think only at the institutional level
(small ball). Mendeley, CiteULike, delicious all are subject based
bibliographies, but with a pan-discipline user base, you start to get the
scale and expertise needed.
Does anybody bother to mine their course reading lists or reserves for
discipline based relevance? This is basically what the project I work on
(Talis' Aspire -- http://www.talisaspire.com/ </product_plug>) does. There
is all kinds of semantic relevance embedded in the ways scholars actually
use materials.
It's also very similar to my (unrealized) vision of the Communicat (
http://journal.code4lib.org/articles/24).
Certainly there are other possible ways to create some of these
relationships (circ data, services like Ex Libris' bX, etc.).
Yes, there is little possibility for a library to accomplish a
bibliography-like experience on it's own. I think that's exactly the
mindset we need to get out of, however.
-Ross.
> In the 21st century (heck, even in the 20th century), the point of the
field of 'information retrieval', building systems to answer user queries,
is trying to make a system that can start from a large corpus, and assemble
what we could call a 'bibliography', a subset of that corpus matching the
user's query. Matching on full text if we have it, matching on controlled
vocabulary if we have it, etc.
>
> Of course, this assembled subset is not going to be as good as a human
expert created bibliography. It's really disrespectful to call it a
'bibliography' at all, except by analogy.
>
> But the idea that there will be human expert maintained bibliographies
created and kept up to date for every possible research topic a user may be
interested in -- is simply a weird luddite fantasy.
>
> Jonathan
>
Received on Thu Aug 11 2011 - 20:59:38 EDT