Classifying Web Sites (fwd)

[Barbara Stewart]

> 
>  Forgive the ravings of a cataloger ;-)
> 
> Why can't a program be written to assist the web page developer in web 
> page classification?  It could contain, for example, the LC Subject 
> Headings, Dewey Decimal, NLM Classification, other thesauri of your 
> choice....Questions would be formulated according to number of uses of a 
> term.  For example, let's use the term "Rock". The program would ask 
> something like "You have used the term "rock" 15 times in this document.  
> Are you discussing : stone?
>                      rock and roll?
>                      rock candy?
>                      the motion of moving from side to side?
>                      rock climbing?
>                      rock gardens?
>                      rock lobsters?
> 
> etc. etc. Say, for example, we choose rock lobsters. It would ask:
> Are you discussing: the lobster industry?
>                     cookery?
>                     shellfish in general?
> 
> and so on. Just like one of those choose-your-own-ending adventure 
> stories, at the end we would have a selection of subject headings 
> classifying our page.  Of course this would not work with terms that did 
> not appear frequently within the document.  But for those terms, we could 
> choose the broad classification area, and be prompted into narrowing down 
> our topic, as Anne Callery discussed.
> 
> Our product could be called AutoClass, WebClass, or something similar. 
> 
> Any takers?
> 
> 
> -- 
> Barbara Stewart, Latin American  Cataloger
> W.E.B. Du Bois Library, Univ. of Massachusetts, Amherst
> Amherst, MA 01003
> (413)545-2728  Fax:(413)545-6494
> stew at library.umass.edu
> 


-- 
Barbara Stewart, Latin American  Cataloger
W.E.B. Du Bois Library, Univ. of Massachusetts, Amherst
Amherst, MA 01003
(413)545-2728  Fax:(413)545-6494
stew at library.umass.edu