- From: Charles McCathieNevile <charles@w3.org>
- Date: Sun, 2 May 2004 15:00:00 -0400 (EDT)
- To: Joe Clark <joeclark@joeclark.org>
- Cc: WAI-GL <w3c-wai-gl@w3.org>
(I think this issue is related to the neccessity of eschewing sesquipedalianism, by the way) On Sun, 2 May 2004, Joe Clark wrote: >> Issue 700 - Examples of ambiguous contracted words? > >Many English uses of "it's." But what's the big deal? Who cares? And where >is the *evidence* (as opposed to supposition) that this harms any disabled >person? Where there is no ambiguity the above question obviously doesn't apply. (I believe that your example is a different case - that of people writing incorrectly - see below). I'm not so sure that there is a case to be made about contractions beyond that for general ambiguity, except to remind people that many contractions are difficult to undrstand, although they are a common device in modern manguage - especially acronyms, and abbreviations used by some people in a wider technical community who might not understand them. (eg see mins of telecons from 2Q-4Q 2002). There is circumstantial evidence in the care taken by organisations like Canon, Boeing, and most military services to avoid ambiguity, because people who don't have perfect language skills, in particular (what some advocates would call situational disability also applies here) can make mistakes. In some cases the harm is limited to taking longer to understand what is written - not an absolute barrier, but not allowing people to work as equals either. In others, there are real consequences. Ambiguous language blew up a satellite because people didn't do the reality check of comparing what the units might mean - and although that might seem like rocket science I learned to write units in primary school so as to disambiguate. And the people who didn't understand were rocket scientists. People who don't understand instructions may be killed (talk to Palestinians or Iraqis - some of them can quote precise examples). They can make everyday mistakes, like voting for the wrong person (according to their actual desires), or pqying for products they didn't want. I still don't think we are ready to think about how to put priorities on the requirements we have, since we are missing too many details of them, and these may not turn out to be the highetst priority items, nor the lowest. >Homographs and polysemic words are features of many languages. The Web >Content Accessibility Guidelines 2.0 will be laughed out of town if it >even flirts with the idea of forcing us to use markup like <span title="it >has">it's</span>. Do you have 5 examples of "it's" which mean "it has", please? Correct usage is unambiguous - "it's" is a contraction of "it is" (in the way that "can't" is a contraction of "can not" - as it happens a contraction which, when spoken in many american accents, I find completely ambiguous since I often cannot distinguish it from "can", its opposite. In speech I actually ask people to precise which they mean, but a computer can be told to do it once and expected to get it right thereafter, which is handy). Perhaps you meant to suggest that it was often incorrectly used in place of "its". and you think we should ask people to identify things they did incorrectly. In that case I disagree, because I think we should ask them to fix them, where in the case of language the primary goal is providing an identifiable meaning for them. For unambiguous words used correctly this is pretty straightforward - software like Babylon (1) has been doing this for ages. Services like http://www.confusingwords.com/index.php?word=its&search=Find can be developed further to make it easier for authors to get things right, so that such systems actually can work. I think it is worth getting experts in language to analyse these questions, as well as people who can identify what actual problems arise. >And I wouldn't advise people to bring up the similar case of adding nikud >to Hebrew. No, I wouldn't have expected you to. But since you did bring it up, repeated discussions over the last year with people who use screen readers and similar technologies in arabic, hebrew and japanese has suggested to me that these three languages are in fact important examples of where there can be requirements to disambiguate words beyond the level of "common practice" (which is after all a completely unspecified concept, very difficult to test even for judges - the only people I know who are regularly called on to make such tests). While writing all Japanese in Kana would be considered excessive, there is clearly a need in a number of cases to disambiguate certain words. If it comes to that, while it is common in everyday typed french to omit many of the accents, there are certain cases where this can cause problematic ambiguities. As with english, I suggest we ask experts in the other relevant languages to complete the details, based on what we can tell them about the actual difficulties that arise, and why. cheers Chaals
Received on Sunday, 2 May 2004 15:00:01 UTC