Quote:
Originally Posted by aleyx
I'm not sure about the HTML5 data-* attributes in EPUB; are they valid? A quick search doesn't make me optimistic, but I'm not a EPUB guru.
N.
|
Hmmm. That's idea... [goes away and tries it]...
Epubcheck doesn't like data-origurl as an img attr. It reports:
attribute "data-origurl" not allowed here; expected attribute "dir", "height", "id", "ismap", "lang", "longdesc", "style", "title", "usemap", "width" or "xml:lang"
I'm not dead sure that epubcheck not liking it is a deal killer, but it does point out what the problem is:
EPUB was defined during that period of time between HTML4 and HTML5 when people were honestly expecting XHTML to become the next web standard. So EPUB specifies it's contents should be XHTML. XHTML is a subset of XML and considerably less forgiving that HTML4 or HTML5. In theory, one could add an xmlscheme to extend img to add another attribute. In practice, it would have to be tested on every book reader out there.
So while using longdesc like this is arguably a violation of standard, it is an attribute that is documented and expected to be a URL by the ereaders already in the field.
(And FYI, while longdesc is remove from HTML5, it is
still valid in XHTML. Probably because, last I heard, the XHTML working groups had all closed.)
I appreciate the suggestion, but I think it's more practical to keep using longdesc right now.