Tuesday, March 31st, 2009
Anne van Kesteren has posted on the specs that are moving out of the large HTML5 mega spec, and into their own little ditties:
The current state of things is that the following editor drafts are separated out from HTML5:
- Server-Sent Events: Defines the
- Web Storage: Defines the database, session, and persistent storage API.
- The Web Sockets API, The Web Socket protocol: Defines the
Meanwhile DanC created an initial draft that will lead to splitting out the URL section as detailed in my last post. I did some testing on CSS URL handling and found out it uses exactly the same algorithm as HTML5 in user agents except that the URL character encoding is always set to UTF-8, just like it works for
XMLHttpRequest. In light of this it would make sense to simply revise the IRI specification to make more sense, but there are not enough proponents for that apparently as it would affect e.g. Atom. I guess creating a separate IRI specification (which is what the HTML5 URL handling is) creates less theoretical issues… I wonder though whether anyone has actually tested Atom clients. Do they really refuse to handle spaces in URLs? Do they really do the right thing per the IRI specification and normalize URLs when the encoding of the feed is not a Unicode encoding? Do any clients that handle IRIs do that properly? I doubt it.
Posted by Dion Almaer at 3:57 am