Talk:SVG in text-html 2009

From SVG
Jump to: navigation, search
  1. There's a discussion on polyglot HTML documents that could apply to SVG as well, that is: to produce a polyglot document one has to validate both against HTML and XHTML. This could be used for SVG as well, and would be simple to implement in validators, simply reparse as XHTML+SVG.
  2. Can SVG going forward use only lowercase attribute- and elementnames? (Touches upon the issue of case-fixup tables)
  3. Need to clarify the point "The SVG WG would like to know the rationale for including certain HTML tags that break out of foreign content mode." unless we want an answer along the lines of "to not break pages with bogus <svg> tags, as found by a well-known study"
  4. There are SVG images with xmlns:xlink="&ns_XLink;" commonly produced by Illustrator, do we want to break those? Relates to the point about xmlns:xlink and breaking on bogus values.
  5. Should we consider having svg:title parsed as RCDATA, like the html:title?
  6. There may be issues with adding encoding detection for SVG documents, as in it may break some non-svg content.
  7. If the SVG element is the root element as we propose, how would HTML content which comes after the SVG image be handled?

The word "polyglot" seems wrong to me. The word generally means "*contains* more than one language", not "simultaneously *is* both one language and yet also another". E.g. a polyglot book contains passages of text in different languages, but any given passage is not simultaneously both one language and yet also another. My natural thought on encountering the term "polyglot document" was that it could be an HTML document with some CSS in it, or with some JavaScript in it - there is then "more than one language" in such a document. To me "polyglot" did not imply "the whole document simultaneously *is* (validates as) more than one language". I'd suggest "protean document" as a better term, since "protean" means "takes on different forms" (coming from Proteus, the greek god who could take on different forms at will), and that seems apt since such documents can be served and take on the form of two different content types. -jwatt