Hi Richard, We believe the existing algorithm presented in the domain section already supports leaving the target part of the mapping unspecified, see: "STEP 3-1-2-5-2. Else (if no mapping is found): Add the string (in its original cases) to the result string." So the behaviour is: if the target of the mapping is left unspecified, the source domain string is copied to the target. As with all the target domain values, it is then up to the consuming processor to decide how to interpret and use the string provided for the target of the mapping. So we don't propose making any changes to the specifications in relation to this issue. Does that address you comment to your satisfaction? Regards Dave On 16/01/2013 19:32, Richard Ishida wrote: > 8.9 Domain > http://www.w3.org/TR/2012/WD-its20-20121206/#domain > > Is it possible to map domains to nothing? For example, in > > domainMapping="automotive '', medical medicine, 'criminal law' law, > 'property law' law" > > automotive is just ignored. I don't really know whether this makes > sense, but it seems that if you are dealing with keywords, such as in > the HTML5 spec > > <meta name="keywords" content="british,type > face,font,fonts,highway,highways"> > > it seems that you might want to exclude a bunch of keywords in order > to home in on the key domain information. That might be easier to do > in that way, rather than in mapping them all to a different word. > > > [comment not reviewed by i18n WG] >
Received on Thursday, 24 January 2013 13:38:27 UTC