Accept this agenda?

-> http://www.w3.org/XML/XProc/2010/05/27-agenda

Accepted.

Accept minutes from the previous meeting?

-> http://www.w3.org/XML/XProc/2010/04/15-minutes

Accepted.

Next meeting: telcon, 3 June 2010?

Paul gives regrets for 3 June

Administrivia

Yay us! XProc is a W3C Recommendation!

Review comments on XML processor profiles

-> http://www.w3.org/XML/XProc/2010/05/wd-comments/

Comment 1, white space handling

Norm: I'm inclined to agree with David that it would be nice, but I'm not sure what we can say.

Henry: Yes, there's a constant grumbling about whitespace
... Can a non-validating but doctype reading processor notice what elements have element only content and ignore whitespace?

Norm: No, I don't think that's conformant.
... From 2.10 in the XML Rec: An XML processor MUST always pass all characters in a document that are not markup through to the application.
... As far as I can tell, we don't give any processor any leeway to discard whitespace.

<ht> http://www.w3.org/TR/xml/#sec-white-space

Henry: Is it conformant to the XML specification for a non-validating processor to report element-content-whitespace? I see nothing that forbids it.

Alex: So does Saxon throw away element content whitespace?

Norm: Yes, I think it does if you perform DTD validation. But that's not the XProc default and it only applies to DTD validation.
... I'm not sure what we can do to help.

Henry: A significant goal of our spec is to improve interoperability. David points out that we don't say.

Norm: Is a minimal processor or a basic processor allowed to do DTD validation?

Henry: We haven't answered that question.
... Another version of the question is: does the result of processing with a basic processor include attribute type information?
... Do we really want to sign up to what I said before about interoperable infoset or are we just setting a lower bound.
... So we could say what properties you will get, but you might get more. That's a generalization of the question.

Norm: So my intuition is to say that you can't do validation.

Henry: But the XML spec doesn't classify processors tightly enough for us to do that.

Norm: True, and the whole point about reading external declarations is so that we get some attribute types.

Vojtech: If we follow Henry's idea of only specifying a lower bound, then aren't we done?

Some rambling discussion of the issues

Henry: If we want to answer the question, we will have to be much more careful about what is conveyed. This will require a careful reading of the XML spec.
... We'll have to take a stand on every optional feature for non-validating parsers in the XML spec.
... The two we've thought of so far are element content whitespace and attribute types.

Norm expresses concern about what parser implementors will do if we attempt to specify specific answers to a bunch of detailed questions.

Alex: What happens with web browsers today?

Alex volunteers to look at what Webkit does

<scribe> ACTION: Alex to investigate what Webkit does [recorded in http://www.w3.org/2010/05/27-xproc-minutes.html#action01]

<scribe> ACTION: Henry to ask Richard what rxp does [recorded in http://www.w3.org/2010/05/27-xproc-minutes.html#action02]

Comment 2, XSLT media type

Henry: I read the definition of the element-content-whitespace property in the Infoset spec as being very carefully worded to allow the possibility that this property could be set through a process other than validation.

Norm: I'm happy to fix this in our example. Lots of folks use text/xsl, but application/xslt+xml is what's registered.

Alex: Yes.

Norm: Objections?

None heard.

<scribe> ACTION: Henry to change text/xsl in the spec. [recorded in http://www.w3.org/2010/05/27-xproc-minutes.html#action03]

Comment 3, XML Base processing

Alex: I'm not sure I understand what the objection is.

Henry: I think all we need to say is that we don't mean anything more than XML Base, but that by depending on our profiles, you must use XML base. You could get the same effect by saying you conform to XML Base, but you don't have to.

Alex: It's a layer cake, I think all we're missing is a refernece to the infoset.

Henry: I think we made an intentional decision not to put square brackets around base URI.
... the XML base spec does not refer to the infoset spec. It doesn't put square brackets around the phrase base URI.
... And that's why we quite consciously didn't introduce a reference to infoset here.
... I think we just need a terminology section like XML Base that says what we mean by "base URI"

Norm: That seems reasonable; let's do that and see if it satisfies the commenter.

<scribe> ACTION: Henry to propose the addition of a terminology section and reply to the commenter. [recorded in http://www.w3.org/2010/05/27-xproc-minutes.html#action04]

Any other business?

None heard.

Adjourned.