Anne van Kesteren wrote: > > Domel wrote: > >> Many case. Two examples: >> 1. Selectros in CSS. Now selectors can only be element like * , E, E >> F, E > F and pseudo-class/pseudo-elements like: E:first-child. But we >> can't matche any E element ID. Atom 0.8 support xml:lang so support >> also E:lang(c) but there aren't any id attributes to CSS 's E#myid > > > Perhaps I should have been more specific. What is the use case of > styling Atom? > > >> 2. XPointers - point some parto of document in URI/IRI for example: >> http://example.org/atom.xml#myid > > > Also, what is the use case for this. What should it actually when > given to a feed reader. > > Note also that you can safely add the xml:id attribute to any element > without causing any harm. It might be that the feed validator dislikes > it, but that should't be a problem.
I concur. XML has many useful features. The Atom spec should only cover ones that are: 1) likely to be useful in a significant portion of cases (the use cases) 2) require clarification in the context of Atom If you look at section 2, where xml:lang and xml:base are mentioned, you can see that they require some clarification (base MUSt be used to resolve relative URLs, lang has some context and inheritence issues). -Nikolas 'Atrus' Coukouma
