[Web4lib] Logging spelling errors

[K.G. Schneider]

Yes, though that would still skew it towards other sites' misspellings. In
fact I wouldn't feel we had done wrong by this eval simply to use those
misspellings as-is. I guess I keep thinking that if it weren't too all-fired
complex, it would be nice to keep it local. 

Karen G. Schneider
kgs at bluehighways.com

________________________________________
From: Richard Wiggins [mailto:richard.wiggins at gmail.com] 
Sent: Monday, February 06, 2006 10:04 AM
To: kgs at bluehighways.com
Cc: Web4Lib; lita-l at ala.org
Subject: Re: [Web4lib] Logging spelling errors

There are lists of common misspellings out there, so you could build your
list of top N search terms, put into Access, and do a database join with the
list of misspelled words.  
 
We do include common misspellings in our Best Bets database at Michigan
State.  My favorite one is:
 
   libary
 
/rich

 
On 2/6/06, K.G. Schneider <kgs at bluehighways.com> wrote: 
As part of a larger project evaluating search engines, we plan to evaluate
the SEs using common spelling errors entered by users on our site. 

Our informal plan is to use the top query reports we're ginning up, and skim
these reports for spelling errors. But we're using aspell as our
spell-checker and wondered if anyone had come up with a way to log 
spell-check errors in a way that allowed you to produce reports limited to
misspelled terms (which could of course then be sorted by frequency,
assuming people are predictable in some of their misspellings). I can think 
of all kinds of refinements and variations on that, but basically a list of
common misspellings is what we need.

N.b. yup, I'm familiar with and follow the aspell list, and yup, web4lib and
lita-l are still where I want to ask this question. 

Karen G. Schneider
kgs at bluehighways.com

_______________________________________________
Web4lib mailing list
Web4lib at webjunction.org 
http://lists.webjunction.org/web4lib/