Corpora: web-search

[Janne Bondi Johannessen]

The web-search option presented at  http://webcorp.connect.org.uk/ is
indeed useful. I tried a search for "subletted", and got 31 hits (with the
AltaVista search option), nicely presented in a concordance. But I wonder
why it didn't give 91 hits, which is the number AltaVista found for me when
I used their search engine directly. I also tried the Metacrawler, which,
surprisingly, gave only 26 hits.

Since this is only a prototype, I would like to suggest two improvements:
1. Include as an option the search engine FAST,
http://www.ussc.alltheweb.com/, which gave 111 hits (20 more than Alta
Vista)
2. Display the web-address-source for each hit.

Janne Bondi Johannessen.

---------------------------------------------------------------------------
Professor Janne Bondi Johannessen     Tlf: + 47-22 85 68 14
Tekstlaboratoriet                     E-post: jannebj at hedda.uio.no
Institutt for lingvistiske fag        Faks: +47-22 85 69 19
Universitetet i Oslo                  Internett: http://www.hf.uio.no/tekstlab
P.b 1102 Blindern
0317 Oslo, Norway
---------------------------------------------------------------------------