Which topic on which list

At 09:08 AM 2001-02-01 -0500, you wrote:
>I suggest that further discussion of this idea take place in the WAI ER
>working group and that we drop the other groups from the list, since they can
>always read the archives..
>

AG::

Let me suggest the following deal:

Integrating an Internet thesaurus or word database into spell checking is
an ER
topic.

The closeness metric that Lisa referred to is deeper than that and I am not
sanguine about doing it with by-extension data in RDF.  But I changed the
subject line to pursue that in PF.

Is ER going to take charge of the text-scrubbing scenario?  This is developing
a pattern of practice for pronounceable text.  Text in this class would have
been checked and pronunciations not intended as indicated in the word database
used for checking would be marked with sounds or pronunciation aids.  This is
an industrial strength definition of "ready for SSML" texts.  The text would
then be marked that it had been so scrubbed.  This means that the
un-pronounced
terms can be trusted (better, not absolutely) to be pronounced as intended
when
processed by the algorithms in the text to speech subsystem.

The markup could be as simple as borrowing from SSML by namespace, I don't
know, but the process needs to be defined and the "this process was applied"
semantic captured somewhere.  That is semantics shared with EARL, where the
evaluation report is in part saying "this evaluation method was applied."

Al

>One useful feature of such a tool would be that it can provide definitions
>along with words.
>
>Cheers
>
>Charles
>
>On Thu, 1 Feb 2001, Lisa Seeman wrote:
>
>  I have once wrote algorithm for helping dyslexics/LD with spell checking.
>
>  In principle, a spellchecker suggests words by mapping one spelling as
close
>  to another. As LD use a different mechanism for spelling, the mapping is
>  often off.
>
>  Practical ramifications: A spell checker will often not suggest the correct
>  word. Homonyms (which witch is which?) or rule based misspellings are
>  undetected.
>  (When "pluralizing" I randomly choose between using a "s" and an "es". I
>  once was writing a memo about test procedures and chouse the wrong way to
>  pluralize the word "test").
>
>  Normal spell checking algorithms do not cut it.
>
>  However rather then insisting that everyone uses my algorithm (which I
>  should unload sometime anyway) a simple Tesaurus, or definitions facility,
>  would help a lot.
>
>  A Tesaurus that can be used backward, so that I can find the word I want by
>  typing in a  word with a similar meaning, would be even better.
>
>  Could, or how could this be integrated into guidelines.
>
>  Could the whole algorithm (quite a nice bit of c code, but can be
rewritten)
>  be linked.
>   through RDF?
>
>  No, I am not taking this on as an action item. My RDF is not up to it.
>
>  .
>
>  -----Original Message-----
>  From: Charles McCathieNevile <charles@w3.org>
>  To: WAI PF group <w3c-wai-pf@w3.org>
>  Date: Thursday, February 01, 2001 7:40 AM
>  Subject: another use case for RDF
>
>
>  >Well, this is hardly new, but it probably needs to be said a lot to
make it
>  >stick. It's forwarded from teh RDF Interest Group. Since Sean is the
>  >perpetrator, he has likely seen it already and can be clearer about the
>  >goals, but the basic idea is dealing with spelling mistakes. I have a
>  similar
>  >point about dealing with illustration of textual content, which can be
done
>  >in similar ways (except we need imagenet to match wordnet :)
>  >
>  >cheers
>  >
>  >Charles
>  >
>  >--
>  >Charles McCathieNevile   
<http://www.w3.org/People/Charles>http://www.w3.org/People/Charles  phone: +61
409
>  134 136
>  >W3C Web Accessibility Initiative    
<http://www.w3.org/WAI>http://www.w3.org/WAI    fax: +1 617
>  258 5999
>  >Location: I-cubed, 110 Victoria Street, Carlton VIC 3053, Australia
>  >(or W3C INRIA, Route des Lucioles, BP 93, 06902 Sophia Antipolis Cedex,
>  France)
>  >
>  >---------- Forwarded message ----------
>  >Date: Wed, 31 Jan 2001 21:52:11 -0000
>  >From: Sean B. Palmer <sean@mysterylights.com>
>  >To: Dan Brickley <Daniel.Brickley@bristol.ac.uk>,
>  >     Graham Klyne <GK-lists@ninebynine.org>,
>  >     William Loughborough <love26@gorge.net>
>  >Cc: www-rdf-interest <www-rdf-interest@w3.org>
>  >Subject: Re: RDF use case: Extending and Querying RSS channels
>  >Resent-Date: Wed, 31 Jan 2001 16:53:52 -0500 (EST)
>  >Resent-From: www-rdf-interest@w3.org
>  >
>  >""" > >Yes, I stumbled across that earlier today, considentally.
>  >> "Spell Check Serendipity" scores again!
>  >
>  >Easily fixable (in N3... of course!).
>  >"""
>  >
>  >@prefix wn:
<<http://xmlns.com/wordnet/1.6/>http://xmlns.com/wordnet/1.6/> .
>  >@prefix :
> 
><<http://lists.w3.org/Archives/Public/www-rdf-interest/2001Jan/0249.html>ht
tp://lists.w3.org/Archives/Public/www-rdf-interest/2001Jan/0249.html> .
>  >:considentally wn:word "considentally" .
>  >:coincidentally wn:word "coincidentally" .
>  >
>  >:considentally = :coincidentally .
>  >
>  >"""
>  >hmmmm..... maybe that should be:-
>  ><> :inContextOf
> 
><<http://lists.w3.org/Archives/Public/www-rdf-interest/2001Jan/0249.html>ht
tp://lists.w3.org/Archives/Public/www-rdf-interest/2001Jan/0249.html> .
>  >
>  >or something?
>  >
>  >--
>  >Kindest Regards,
>  >Sean B. Palmer
>  >@prefix : <<http://webns.net/roughterms/>http://webns.net/roughterms/> .
>  >[ :name "Sean B. Palmer" ] :hasHomepage
<<http://infomesh.net/sbp/>http://infomesh.net/sbp/> .
>  >"""
>  >
>  >
>
>
>-- 
>Charles McCathieNevile   
<http://www.w3.org/People/Charles>http://www.w3.org/People/Charles  phone: +61
409 134 136
>W3C Web Accessibility Initiative    
<http://www.w3.org/WAI>http://www.w3.org/WAI    fax: +1 617 258 5999
>Location: I-cubed, 110 Victoria Street, Carlton VIC 3053, Australia
>(or W3C INRIA, Route des Lucioles, BP 93, 06902 Sophia Antipolis Cedex,
France)
>

Received on Thursday, 1 February 2001 09:58:02 UTC