Closed michaelmulley closed 11 years ago
Possible drawback: could that be an issue for someone who would like to parse the links with an atom lib?
I don't think so, because an Open511 document isn't an Atom or AtomPub file -- we use one tag, but we're not compliant with the rest of the spec. (There may well have been reuse value in making Open511 an extension to Atom, and using AtomPub, but I think we're too far along to reconsider that.)
And in any case, there isn't really anything to parse that requires more than short XPath expression. If you want to find all the URLs to related links, you don't use an Atom library, you use XPath: link[@rel=related]/@href
Changed
Proposal: Instead of using
<atom:link>
tags in XML, use the tag within our namespace (<link>
). The tag otherwise works the same.Benefit: we can write an XML Schema 1.0 schema for the format, though it wouldn't work for multilingual content; an XML Schema 1.1 schema would work for everything.
In practice, I can't think of any drawbacks to this change, or any ways in which having the
link
tag in the Atom namespace would actually help real-world clients. (We're not using the rest of the AtomPub spec...)