- From: David Carlisle <davidc@nag.co.uk>
- Date: Wed, 9 Apr 2008 19:09:05 +0100
- To: oedipus@hicom.net
- CC: wai-xtech@w3.org
> if the creation of a homogeneous set of XML entity names > across vocabularies has ramifications for assistive technologies... If the applications are using an XML parser (or html parser as the case may be) then the use or not of entity references should be completely transparent. Entity references are an authoring convention and an application isn't supposed to treat a reference to a character any differently to the referenced character, and an xml parser will (typically) replace the entity reference by the character before an application accesses the data. > does an AT need to be able to identify " as a quote, Numeric references are not entity references and not affected by the xml-entity-names-draft, they always9in xml and html4 or 5) refer to Unicode code points, so the same comments apply to numeric references as character references, they should be replaced by the parser before the application sees the data " and " are the same thing in element content. > is it merely a case of ATs needing to fully support unicode, or at least > the discrete set outlined in: > http://www.w3.org/TR/xml-entity-names/ they need to do that in any case, whatever the choice of the mapping of the entity name phi to U+03D5 or U+03C6, the author has the choice (and is encouraged) to use the character directly, so the application needs to be able to accept these characters, not just the name "phi". Of course Unicode has a lot of characters and any system will support some better than others. I reference those blocks in the working draft to give a rough idea of which characters are most likely to appear as symbols in scientific documents, but a more canonical source of that information is Unicode report 25 http://unicode.org/reports/tr25/ David
Received on Wednesday, 9 April 2008 18:10:54 UTC