- From: Charles McCathieNevile <charles@w3.org>
- Date: Sat, 9 Aug 2003 16:49:39 -0400 (EDT)
- To: Jason White <jasonw@ariel.ucs.unimelb.edu.au>
- Cc: lisa seeman <seeman@netvision.net.il>, WAI GL <w3c-wai-gl@w3.org>
Interestingly it is easy to test whether something has complied by machine, since the process essentially involves testing that the content is entirely the things that are (and here we get vague for today) widely understood language. Describing a language in sufficient detail to do this is not a trivial task - for small and restricted languages like the Yolngu Matha group from Northern Autralia it could perhaps be done in a few years - for the Quinkan language it is not clear that the remaining native speakers will live long enough to permit it to be done. On the other hand, large and fluid languages like english present a challenge of keeping the information up to date. There are two approaches to making a dictionary (and grammar). One is that typified by the Oxford English Dictionary, which is essentially empirical - it seeks to record the language as it is in fact used. The second is to write down how the language "should" be used - attempting to define good and bad usage. In countries such as France, Spain, the Vatican, and Iceland, there are official bodies which do this. It strikes me that what is important for WCAG is the empirical approach to understanding what really are commonly understood terms, rather than knwoing what "ought to be", which is perhaps an easy way to get 80% accuracy but not capable of providing better. (Note that taking into account further issues such as unambiguous use of terms complicates further in some languages. Others, like Esperanto, make that impossible...) cheers Chaals On Sat, 9 Aug 2003, Jason White wrote: >In fact it might even be machine testable if there is a generative >grammar for the required syntax and a list of permissible vocabulary; >and before anybody asks, it is a fundamental insight of modern >linguistics, as I understand it, that the syntax of every natural >language can be described by a generative grammar. > Charles McCathieNevile http://www.w3.org/People/Charles tel: +61 409 134 136 SWAD-E http://www.w3.org/2001/sw/Europe fax(france): +33 4 92 38 78 22 Post: 21 Mitchell street, FOOTSCRAY Vic 3011, Australia or W3C, 2004 Route des Lucioles, 06902 Sophia Antipolis Cedex, France
Received on Saturday, 9 August 2003 16:49:40 UTC