>> Is your complaint against including ZIP-package-relative URLs in >> the original content (whether it is 'widget:' or 'zip:'), or is >> your complaint against using a 'zip:' protocol in the original >> content instead of a 'widget:' protocol? > > As far as authoring is concerned, I think they're (almost) equally > bad.
Authors would not be allowed to use the scheme. > Quite generally, it would appear that using a URI scheme that's > specific to the packaging format basically leads to a layering > violation -- so far, the widget's DOM layer doesn't need to have a > clue whether it was packaged with zip (as the current spec says), > not packaged at all (as in, installed in a widget engine that just > unpacks things into the file system), or whatever else. So you say that what relative and absolute URIs resolve to internally is an implementation detail? -- Marcos Caceres http://datadriven.com.au
