#494 xml:base

closed-accepted
nobody
5
2013-01-19
2013-01-06
No

I think this is a bug, though I'm not sure where exactly.

According to the Guidelines, @xml:base follows the W3C spec, http://www.w3.org/TR/xmlbase/. That sounds good and right.

TEI says the datatype for @xml:base is data.pointer. But by the spec the value for an individual @xml:base value need not itself be a valid URI. It can be a piece of a URI. A full URI gets hierarchically assembled from the individual parts in each of the xml:base values in the tree.

So an individual @xml:base in a TEI document need not be a data.pointer, contra the spec. Only the complete string assembled from such pieces need be.

I think the datatype for @xml:base should be changed and a note added somewhere saying that all pieces must form a valid URI.

Discussion

  • Sebastian Rahtz

    Sebastian Rahtz - 2013-01-06

    isn't a partial/relative URI still a URI? how do you tell the difference?

     
  • John P. McCaskey

    Well, I didn't think a path segment was a valid URI.

    But I've just read through RFC 3986 and now believe I was wrong and Sebastian is right. A segment, all by itself, is a valid (relative) URI.

    I'm still a little uncertain that saying xml:base must be a data.pointer and saying xml:base must be as defined in http://www.w3.org/TR/xmlbase/ are equivalent statements.

    But if I learn more and still think the Guidelines need tweaking, I'll reenter a bug report with more detail.

    As it is, you can reject this one.

     
  • John P. McCaskey

    I did a little more research and got an answer back from Michael Kay on xml-dev on the one case that was bugging me.

    Something that threw me was use of the phrase "a single URI pointer" in the spec for data.pointer. "URI pointer" is not a term used in 3986 or the xml:base spec.

    I suggest the wording for data.pointer be changed from

    defines the range of attribute values used to provide a single URI pointer to any other resource, either within the current document or elsewhere.

    to

    defines the range of attribute values used to provide a single URI, absolute or relative, that points to any other resource, either within the current document or elsewhere.

    I also suggest "RFC 3986" in the notes be made a link, just as "RFC 3987" is in the succeeding sentence.

     
  • Sebastian Rahtz

    Sebastian Rahtz - 2013-01-06

    i have made 3986 into a link.

    I'll leave the wording change to others to mull over.

     
  • Lou Burnard

    Lou Burnard - 2013-01-19

    The wording looks fine and I've changed it as suggested, modulo the phrase "that points to any" which I think reads more fluently as "pointing to some". Apologies for not having noticed this in time for the recent release, but the change has now been made (rev 11483)

     
  • Lou Burnard

    Lou Burnard - 2013-01-19
    • status: open --> closed-accepted
     

Get latest updates about Open Source Projects, Conferences and News.

Sign up for the SourceForge newsletter:





No, thanks