See also: IRC log
Henry and Norm gave regrets for 10 Dec; Paul is at risk; Murray possibly. Let's cancel 10 Dec. See you all on 17 Dec.
Henry: This seems to me to be
about the question of what is the error. The answer is there's
been a static error since day one.
... What is the error for p:banana in p:for-each
Norm: For ordinary behavior, I think the answer is clear. I think this is about fowards-compatible behavior.
<ht> or my:zoolander in p:for-each
Vojtech: It's a static error because it's an unknown element or a dynamic error if it's an unknown step.
Norm: It's a static error in the my:zoolander case because you have to have imported a decl for it.
Henry: You have to use use-when.
Vojtech: With respect to instruction-avaialble, you can always use the version.
Norm: You can check for an available step, but not an available new thing like p:new-serialization.
Vojtech: That would be a new language feature.
Norm: Right, and I think we're willing to live with a little less flexibility there.
Vojtech: Checking for p:new-serialization wouldn't be flexible enough, you might want to check for an attribute for example.
Henry: Basically, you want to say
"use-when this stylesheet will work" and that's just not easy.
You'll have to use version numbers for now.
... Version-available is all you get in V1 and that's going to have to be enough.
<scribe> ACTION: Henry to follow-up with Vasil. [recorded in http://www.w3.org/2009/12/03-xproc-minutes.html#action01]
<scribe> ACTION: Editor to consider whether it's possible to clarify the definitions of atomic and compound steps. [recorded in http://www.w3.org/2009/12/03-xproc-minutes.html#action02]
Vojtech: There were two things. I think no action is necessary.
Norm: Perfect, I'll close that one.
This was discussed at the 12 November meeting. Decision was that no name change was necessary.
Norm: I updated 5.12
Henry: I saw that and concluded that my action had been overtaken.
Norm: Proposal: close as resolved
Norm: Updated 3.9. Is everyone satisfied?
Henry: I'm still concerned about listing only what isn't possible. I had listed the things that are available.
Vojtech: If you use use-when in a pipeline labeled version=1.0, then step-available only sees the 1.0 steps, right?
Vojtech: Which means that during pre-processing you must take version into account.
Norm: Yes, I think that's
... I think the fact that you're in a 1.0 pipeline is something you must know when you evaluate use-when expressions.
Vojtech: So a 1.0 pipeline with a 2.0 declare-step, if the declare-step includes a use-when you have to use 2.0
Norm: Yes, that's probably worth mentioning.
<scribe> ACTION: Norm to clarify that version comes into play in use-when [recorded in http://www.w3.org/2009/12/03-xproc-minutes.html#action03]
Henry: Why doesn't this just work?
Norm: Because the XPath 2 type
system doesn't promote from xs:string to anything
... I think this just a limitation in v1.0
Proposal: We just have to live with it.
Norm: Any objections?
Norm: No new features at this point.
Norm: I think the upshot is only viewport and for-each change the values
<scribe> ACTION: Editor to clarify [recorded in http://www.w3.org/2009/12/03-xproc-minutes.html#action04]
<scribe> ACTION: Alex to investigate and report back [recorded in http://www.w3.org/2009/12/03-xproc-minutes.html#action05]
Vojtech: In making the request, if you say encoding=base64 with no charset, then you just send the decoded byte sequence, right?
Alex: c:body has content-type but no charset
Norm: Right, I think Vojtech's
argument to put it in the content type is compelling.
... Yes, c:data on those steps should be consistent with unescape markup
Henry: I'm happy with new-prefix.
Norm: I propose to add
new-namespace and new-prefix (with appropriate names) to all
the steps that have wrapper element names as options.
... Any objections?
Norm: No new features.
... I think I should have filtered this one out.
Henry: I'm happy with it.
Vojtech: I have one small thing regarding the XQuery step.
<alexmilowski> I'm happy with the most recent draft too.
Vojtech: I think in Santa Clara we concluded that for things like XQuery or text input, we don't inherit any namespace bindings.
Norm: I'd have to review the minutes again.
Norm: We're out of time, I'll take it to email