W3C home > Mailing lists > Public > public-xml-processing-model-wg@w3.org > October 2011

log of today's xproc f2f

From: Grosso, Paul <pgrosso@ptc.com>
Date: Mon, 31 Oct 2011 19:56:36 -0400
Message-ID: <9B2DE9094C827E44988F5ADAA6A2C5DA03EEEA9F@HQ-MAIL9.ptcnet.ptc.com>
To: <public-xml-processing-model-wg@w3.org>
I didn't do anything with rssagent before I logged off irc,
so I don't know what happened to our minutes.  Hopefully
someone will be able to capture it from wherever it is.

But meanwhile just in case, I'm including as the rest of
this email message my locally-saved irc log of all of today's
xproc f2f meeting.



Session Start: Mon Oct 31 09:04:13 2011
Session Ident: #xproc
[09:04] *** Now talking in #xproc
[09:04] *** Topic is 'XProc meets 20 October,
[09:04] *** Set by Norm on Thu Oct 20 07:02:44
[09:04] *** MoZ has joined #xproc
[09:04] *** Zakim has joined #xproc
[09:04] *** RRSAgent has joined #xproc
[09:04] <RRSAgent> logging to http://www.w3.org/2011/10/31-xproc-irc
[09:05] <ht>
[09:05] *** ht changes topic to 'Agenda:
[09:05] <MoZ> Scribe: alexmilowski
[09:10] <alexmilowski> Present: Henry Thompson, Paul Grosso, James
Fuller, Vojtech Toman, Mohamed Zergaoui, ,Cornelia Davis,  Murray
Maloney, Alex Milowski
[09:11] <alexmilowski> Issues list:
[09:14] <alexmilowski> Issue 3: We think this is closed but need to
check with Liam to verifiy.
[09:14] *** Cdavis_ has joined #xproc
[09:16] *** Cdavis_ has quit IRC (Quit: Page closed)
[09:16] <ht> http://www.w3.org/XML/XProc/docs/xml-proc-profiles.html
[09:16] *** Cornelia has joined #xproc
[09:16] *** jfuller has joined #xproc
[09:16] <ht>
[09:19] <alexmilowski> Action: Editors to remove "particularly" clause
in section 5 as this may lead to inferences that we do not want.   See
Henry's brain.
[09:19] * RRSAgent records action 1
[09:24] <alexmilowski> Disccusion of Issue 7: Henry was recalling his
memory of the lead up to taking this to XML Core and how XHTML treats
entity definitions.
[09:25] <alexmilowski> Henry: standalone=yes does not cause an error ...
no difference for a well-formed parser.
[09:26] <alexmilowski> Henry: No need to change the default because it
won't change the behavior of the parsers in use for XHTML in browsers.
[09:27] *** ht has quit IRC (Ping timeout)
[09:29] <alexmilowski> Henry: XHTML5 maps any public identifier to a
pre-defined external subset.
[09:32] <alexmilowski> Paul: What are the datatypes available?
[09:32] <alexmilowski> Henry: Only DTD datatypes.
[09:32] <alexmilowski> Some discussion of attribute definitions,
NMTokens, and tokenization available when parsers encounter definitions.
[09:36] <alexmilowski> Some more digging into what HTML5 says about
doctypes and entity definitions.
[09:40] <alexmilowski> Henry: Core says "there is nothing here" ...
we're done.
[09:40] <alexmilowski> Alex: I agree
[09:40] <alexmilowski> Action: Close issue 7
[09:40] * RRSAgent records action 2
[09:42] <alexmilowski> Action: Henry to try to get agreement from Henri
Sivonen on issue 9.
[09:42] * RRSAgent records action 3
[09:47] <alexmilowski> Discussion on Issue 19
[09:47] <alexmilowski> Paul: Look at paragraph 3 ...
[09:48] <alexmilowski> Action: Recommend the editor use Paul's version
of paragraph 3.
[09:48] * RRSAgent records action 4
[09:48] <alexmilowski> Action: Change "is an attempt to give" to "gives"
[09:48] * RRSAgent records action 5
[09:49] <alexmilowski> Action: Norm needs to suggest how this will be
integrated into the document.  Where does this fit in?
[09:49] * RRSAgent records action 6
[09:50] *** ht has joined #xproc
[09:51] <MoZ> http://www.w3.org/2011/10/31-xproc-irc
[09:52] <ht> HST proposes wrt issue 19 that we relabel section 1 as
'Introduction', push the existing prose down to subsection 1.1
Background, and make Norm's new prose the body text of section 1
[09:54] <alexmilowski> With the actions taken today, we will have closed
all the issues today, hopefully to the satisfaction of the commentators.
[09:54] <alexmilowski> Last issue is issue 8
[09:56] <jfuller>
[09:58] <alexmilowski> Issue 8.2: Address by adding different profiles.
[09:59] <alexmilowski> Issue 8.2: We kept basic and added external
declaration profile.
[09:59] <alexmilowski> Henry: Agreed.
[09:59] <alexmilowski> Working group endorses suggested solution to 8.2
[10:00] <alexmilowski> Issue 8.3: Class V starts to address this.  Do we
need to add a full validating profile?  We've minimally addressed this.
[10:00] *** Cornelia has quit IRC (Ping timeout)
[10:01] <alexmilowski> Issue 8.3: We've also added section 7:
[10:01] <alexmilowski> Working group endorses solution to 8.3
[10:02] <alexmilowski> Issue 8.5: Possibly add a diagram?
[10:04] <alexmilowski> Henry: This should be added to 4.2
[10:05] <alexmilowski> Vojtech: Maybe we want to rename the profile
classes so that they make more sense in the diagram?
[10:07] <alexmilowski> Henry: We need all the classes we have.  Renaming
them may make sense.
[10:07] <alexmilowski> Alex: Maybe we finish the diagram and see what
makes sense.
[10:10] <alexmilowski> Some discussion between Henry & Murray about the
classes & validation.
[10:13] <alexmilowski> Murray: Validating XML processors must read &
process the external declarations... 
[10:13] <alexmilowski> Henry: (Reads the spec saying that as so ...)
[10:14] <alexmilowski> A processor that validates but doesn't read
external declarations isn't a conforming XML processor.
[10:16] <alexmilowski> Henry: (Paraphrasing the discussion) Making
validation optional or required is incoherent against the profiles.
[10:17] <alexmilowski> Alex: I'm feeling uneasy about this.  As a user
you can pick a profile and turn on validation and do the wrong thing.
[10:17] <alexmilowski> Henry: We need to say something about this.
[10:20] <alexmilowski> (Murray is point at the diagram making good
points about validation and profiles.)
[10:26] <alexmilowski> Murray: In the case where you are enabling the
XInclude and validation flag, can we say "it is recommended" or
"required" that you validate after the XInclude?
[10:27] <alexmilowski> Henry: Instead of three, there are only two:
before or after.
[10:27] <alexmilowski> Paul: Isn't that the [status quo].
[10:28] <alexmilowski> Henry: It is coherent to validate first because
you'll get element content whitespace ...
[10:29] *** MoZ has quit IRC (Quit: This computer has gone to sleep)
[10:30] *** Vojtech has quit IRC (Ping timeout)
[10:31] *** jfuller has quit IRC (Ping timeout)
[10:42] * PGrosso Zakim, who is on the phone?
[10:42] *** Disconnected
Session Close: Mon Oct 31 10:42:16 2011

Session Start: Mon Oct 31 10:42:28 2011
Session Ident: #xproc
[10:42] *** Now talking in #xproc
[10:42] *** Topic is 'Agenda:
[10:42] *** Set by ht on Mon Oct 31 08:05:32
[10:46] *** Cornelia has quit IRC (Ping timeout)
[10:48] *** Cornelia has joined #xproc
[10:49] <ht> RRSAgent, logs?
[10:49] <RRSAgent> I'm logging.  Sorry, nothing found for 'logs'
[10:49] <ht> RRSAgent, log?
[10:49] <RRSAgent> I'm logging.  Sorry, nothing found for 'log'
[10:50] *** jfuller has joined #xproc
[10:55] <alexmilowski> Comment: We have "id xml processor profile" but
the profile adds xml:id.  Maybe this should be the "xml:id XML Processor
[10:55] <alexmilowski> Comment: Maybe we should make "XML Processor
Profile" less redundant in the document.
[11:02] <alexmilowski> Paul: On issue 8.5, what are the remaining
[11:03] <alexmilowski> The validation questions relate to 8.3.  We need
to re-open this issue.
[11:04] <alexmilowski> Paul: We can close 8.5 by adding the diagram.
[11:05] <alexmilowski> Henry: It is perfectly valid to provide XML
Schema validation for any of the profiles.  ... it is not the same for
DTD validation.
[11:06] <alexmilowski> Paul: We can close 8.3 and 8.5 and open a new
issue about validation.
[11:06] <alexmilowski> Action: Henry will draft the new issue.
[11:06] * RRSAgent records action 7
[11:07] <ht> New issue: How to expand 7 (and possible earlier bits) to
clarify the distinction between DTD validation and validation in general
[11:07] *** Liam has quit IRC (Client exited)
[11:07] *** Liam has joined #xproc
[11:07] <ht> ... DTD validation is _not_ orthogonal, e.g. Basic+DTD
Validation is not conformant with XML spec
[11:08] <ht> ... but e.g. Schema validation is orthogonal
[11:08] <PGrosso> s/7/section 7 in the draft/
[11:08] <ht> ... Also, expand the discussion of ordering of xinclude and
[11:10] <alexmilowski> Issue 8.7: In profiles external declarations
(2.3) and full (2.4), "reading and processing" versus "processing." 
[11:11] <alexmilowski> Henry: That prose is directly from the XML
specification and I'm reluctant to fix it.
[11:11] <alexmilowski> Henry: [ this text intended to reproduce what the
XML spec says ]
[11:14] <alexmilowski> The link in the profiles document takes you to
the location in the XML specification that has the relevant text.
[11:18] <alexmilowski> Action: Henry will attempt to separate the two
parts of #1 on 2.3/2.4.
[11:18] * RRSAgent records action 8
[11:19] *** MoZ has joined #xproc
[11:22] *** ht has quit IRC (Ping timeout)
[11:22] <alexmilowski> Action: Issue 8.8: Editorial
[11:22] * RRSAgent records action 9
[11:26] <alexmilowski> Henry: Instead of steps necessary, they are steps
"preparatory" .
[11:26] <alexmilowski> The profile steps are not "steps" ...
[11:26] <alexmilowski> Action: Henry will rework the introduction to
section 2.
[11:26] * RRSAgent records action 10
[11:27] <alexmilowski> "Step" is the wrong word throughout the profile
section ... Henry will look at this as well.
[11:28] <alexmilowski> Issue 8.9, action to henry
[11:29] <alexmilowski> Action: Henry to review the use of conformance in
section 4.
[11:29] * RRSAgent records action 11
[11:29] * Zakim excuses himself; his presence no longer seems to be
[11:29] *** Zakim has left #xproc
[11:30] <alexmilowski> Issue 8.10: There have been changes that may have
addressed this.
[11:31] <alexmilowski> Henry: word 'rigid' is still there.
[11:32] <alexmilowski> Action: Henry to soften language in the first
paragraph of Section 1, Background.
[11:32] * RRSAgent records action 12
[11:34] <alexmilowski> Issue 8.12
[11:34] <alexmilowski> Henry: remove "since this specification is not
implementable as such" and this will be fixed.
[11:35] <alexmilowski> Alex: What did the infoset do about this?
[11:38] <alexmilowski> Alex: We use "require" in each of the profile.
[11:38] <alexmilowski> Henry: We define conformance ...
[11:38] <alexmilowski> ...make it be that conformance starts when some
other specification references our specification.
[11:39] <alexmilowski> Henry: It is going to define what it means to
conform to a profile.
[11:40] <alexmilowski> Henry: [ a substantive change to section 6 to
address issue 8.12]
[11:41] <alexmilowski> This specification doesn't have implementations
but it does have specifications that conform to it.
[11:41] <alexmilowski> Action: Henry to change section 6 to address 8.12
[11:41] * RRSAgent records action 13
[11:42] <alexmilowski> Issue 8.14
[11:44] <alexmilowski> Henry: XPath 2 distinguished between
implementation defined and implementation determined.
[11:44] <alexmilowski> [choice vs unspecified]
[11:46] <alexmilowski> implementation dependent vs implementation
[11:46] <alexmilowski> Action: Henry to clarify use of term to address
8.14.  Take suggested fix.
[11:46] * RRSAgent records action 14
[11:47] <alexmilowski> Issue 8.16
[11:47] *** Vojtech has joined #xproc
[11:47] <alexmilowski> The names may change again
[11:47] <alexmilowski> In progress...
[11:49] <alexmilowski> Henry to consider moving the tabulation to the
front of section 3.
[11:51] <alexmilowski> Alex: I like having the class definitions first
so you know what the table is about.
[11:51] <alexmilowski> Henry: It might be more useful to have more
descriptive names: Class A: Items and properties fundamental to all XML
[11:52] <alexmilowski> Alex: Maybe change the class definitions to have
two parts: the description of the class and the requirements on the
[11:55] <alexmilowski> Issue 8.4
[11:56] <alexmilowski> Still open, James is building a list.
[11:56] <alexmilowski> Issue 8.11
[11:58] <alexmilowski> Henry: We can use this for an implementation
[11:59] <alexmilowski> Alex: There is a distinction between the options
and the common use of those options in a product (e.g. Chrome/Safari)
[11:59] <alexmilowski> James is working on this.
[11:59] <alexmilowski> Alex: I volunteer for helping with WebKit et. al.
[11:59] <alexmilowski> Issue 8.6
[12:01] *** Liam has quit IRC (Client exited)
[12:01] *** Liam has joined #xproc
[12:02] <alexmilowski> Henry/Murray: "ID type assignment" language...
[12:07] <alexmilowski> Action: Take the suggestion by using
http://www.w3.org/TR/xml-id/#inform and taking "ID type assignment" and
forcing bullet #1.  See minutes.
[12:07] * RRSAgent records action 15
[12:08] *** Liam has quit IRC (Ping timeout)
[12:09] *** ht has joined #xproc
[12:09] <alexmilowski> "Perform ID type assignment for all xml:id
attributes as required by xml:id 1.0 by setting their attribute type
Infoset property to type ID"
[12:10] <alexmilowski> Issue 8.13
[12:10] <alexmilowski> Section 6 is a start...
[12:11] <ht> RRSAgent, make logs public
[12:11] <RRSAgent> I have made the request, ht
Session Close: Mon Oct 31 12:12:20 2011

Session Start: Mon Oct 31 13:45:02 2011
Session Ident: #xproc
[13:45] *** Now talking in #xproc
[13:45] *** Topic is 'Agenda:
[13:45] *** Set by ht on Mon Oct 31 08:05:32
[13:46] <alexmilowski>
[13:47] <PGrosso> scribe pgrosso
[13:47] <PGrosso> 3 things left from Jim's list
[13:48] <PGrosso> open issues against XProc itself which we need to sort
into Vnext requests and potential errata.
[13:51] <PGrosso> http://www.w3.org/XML/Activity
[13:51] <MoZ> http://www.w3.org/XML/2010/10/xproc-charter
[13:53] *** Cornelia has joined #xproc
[13:55] <PGrosso> Our charter ends the end of this coming January. We
need to decide if we will recharter or just extend.
[14:00] *** ht has quit IRC (Ping timeout)
[14:01] <PGrosso> The abstract says each profile defines a data model,
but that isn't really true. We should consider rewording that.
[14:01] <PGrosso> The profile determines properties that are available
from which to determine a data model.
[14:02] <PGrosso> action to henry: the abstract (and any paragraph in
the Background that is almost a copy of it) needs to be rewritten.
[14:03] *** ht has joined #xproc
[14:05] <PGrosso> We find that the spec uses the term data model all
over the place and perhaps in a fashion that will be confusing to
[14:05] <PGrosso> Jim's terminology section should define the term,
though Alex suggests perhaps we use a different word in most cases.
[14:07] <PGrosso> action to alex: sketch out by tomorrow morning if
possible how we should address the "data model" terminology in the spec.
[14:11] <PGrosso> Perhaps we should add some words to explain why we
picked each of the 4 profiles we did and admit that there could be lots
more so that our choice was somewhat arbitrary although still, we hope,
[14:12] <PGrosso> For example, we believe all browsers implement (at
least) basic and not all browser implement any of the larger profiles.
[14:13] <PGrosso> Our profiles were based on sets of available
properties, not on things like streaming or not or dynamic manipulation
or not.
[14:16] <PGrosso> action to jim: suggest a short rationale for our
picking each of our profiles.
[14:18] <PGrosso> Hey HENRY:  what does "faithful provision" mean?
[14:18] <ht> Where?
[14:19] <PGrosso> In section 2 all over
[14:20] <ht> It means that whatever gets put in a data model does
actually (enable itself to) reconstruct the information defined by the
relevant infoset property
[14:20] <PGrosso> ht:
[14:21] <ht> So, e.g., if the parser builds a datamodel that doesn't
actually discriminate between NMTOKEN and ID is not 'faithfully
provisioning' wrt the attribute type
[14:21] <ht> property
[14:21] <PGrosso> I don't think I understand that use of the word
"provision". Can you give me a synonym?
[14:22] <ht> 'install'
[14:22] <ht> 'install in'
[14:22] <PGrosso> for 8.15, we will accept Michael's suggested fix and
let the editor massage as necessary.
[14:23] * ht checking back out -- using my name will indeed call me over
[14:25] <PGrosso> for 8.17, Alex suggests we add a short sentence or two
about each of xml:id, xml:base, and xinclude to section 2 (perhaps just
the intro to 2 or maybe a new subsection).
[14:25] <PGrosso> [and the WG agrees]
[14:26] <PGrosso> action to Murray: suggest the wording to add about
xml:id, xml:base, and xinclude.
[14:28] <PGrosso> For 8.18, these are all editorial, and we are leaving
their resolution to Jim and Norm.
[14:29] <PGrosso> And that takes us to the end of LC comments.
[14:31] <PGrosso> Section 4.2.3, Vojtech questions whether "Unexpanded
Entity Reference Information Items" should be in there at all because he
doesn't think there is any difference.
[14:31] *** Cornelia has quit IRC (Ping timeout)
[14:34] <PGrosso> Also in section 4, we note that all those "Entirely,
for the same reason" are still confusing and need to be spelled out or
[14:38] <PGrosso> We believe (though we're not positive) that
"Unexpanded Entity Reference Information Items" has to be the same for
the "external declaration" profile and the full profile.
[14:39] <PGrosso> We aren't sure that we understand what happens for
Unexpanded Entity Reference Information Items for either profile, so we
need to re-discuss this with HST.
[14:41] <PGrosso> Vojtech has some editorial comments that he will pass
on to Jim.
[14:46] <alexmilowski>
[14:46] <jfuller> http://www.w3.org/XML/XProc/xproc-candidate-issues/
[14:48] <PGrosso> issue 001 to be filed under Vnext.
[14:48] <PGrosso> Issues 002 and 003 are closed.
[14:48] <PGrosso> Issue 004 is for V.next.
[14:51] <PGrosso> Issue 005 is about conformance for the xproc (and
Vojtech's comment here is about the profile spec), so this goes into the
errata pile.
[14:52] <PGrosso> Issue 006 is for V.next.
[14:54] *** jfuller has quit IRC (Ping timeout)
[14:54] <PGrosso> We believe that issue 007 is a bug in Calabash.
action to Norm to check and confirm.
[14:55] <PGrosso> Issue 008 is an erratum.
[15:00] <PGrosso> Issue 009 is asking that the xproc schema be updated
to include p:template, but p:template is not part of V1, so we wonder if
we can change the schema. Paul doubts it, but thinks that we could add
such a schema to the p:template note. We should discuss this with Norm
and Henry too, but we are leaning toward adding the augmented schema to
the note.
[15:02] <PGrosso> MoZ says that implementors cannot add something in the
p namespace, so they cannot use p:template with the official xproc
[15:02] <PGrosso> Leaving Issue 009 open for discussion.
[15:05] <PGrosso> At least most of Issue 010 is V.next. But there is one
thing that Norm says "I'll put that on the bug list" and we're not sure
what that is, so:
[15:06] <PGrosso> action to Norm: Look at issue 010 and determine what
aspect of it is a bug and report back.
[15:06] *** MoZ has quit IRC (Quit: This computer has gone to sleep)
[15:08] *** Vojtech has quit IRC (Ping timeout)
[15:30] <PGrosso> [break until 15:45]
[15:58] <alexmilowski>
[15:59] <PGrosso> Issue 013 was discussed in the minutes Alex just
posted above.
[16:00] <PGrosso> Norm was given an action to write a proposal for
[16:00] *** MoZ has joined #xproc
[16:00] <PGrosso> And Alex has an action on this issue to do some more
[16:03] <PGrosso> Issue 014 is an erratum.  It requires some
clarification in the spec as outlined in Vojtech's email at
[16:06] <PGrosso> Issue 015 is V.next unless Norm says it's just
closable.  action to Norm to confirm.
[16:07] <PGrosso> Same with Issue 016--V.next with Norm to confirm.
[16:11] <PGrosso> Issue 017 is V.next.
[16:12] <PGrosso> We believe issue 018 is just fyi and is neither an
erratum or A v.next request, so we will just close it.  action to Norm
to confirm.
[16:13] <PGrosso> Issues 019 through 024 are already closed.
[16:14] <PGrosso> Issue 025 is an erratum.  We should clarify that xslt
match patterns are evaluated using the Step xpath context.
[16:14] *** alexmilowski has quit IRC (Quit: alexmilowski)
[16:20] <PGrosso> action to Norm (editor): clarify that xslt match
patterns are evaluated using the step xpath context (to close 025).
[16:22] <PGrosso> action to Norm (editor): Clarify that what Norm asked
about is conformant to close 014.
[16:25] <PGrosso> action to Norm (editor): to correct the obvious bug
outlined in issue 008.
[16:32] <PGrosso> Regarding 009, Paul suggests that we can add to the
p:template WG Note an augmented schema, but we can't pretend that the
augmented schema is the official 1.0 one.
[16:34] <PGrosso> action to Jim: create the augmented schema (that
includes p:template) and augment the WG Note to point to the augmented
[16:36] <PGrosso> So we can close 009 as neither errata nor V.next
(though we'll probably put p:template into V.next) and just address it
with a Second Edition of the WG Note.
[16:38] <PGrosso> That leaves us with 005 on the conformance section of
the xproc spec at http://www.w3.org/TR/xproc/#conformance
[16:42] <PGrosso> action to Jim:  Give a try to finding all conformance
statements throughout the spec and putting references to them in the
conformance section to address issue 005.
[16:50] <PGrosso> meeting adjourned 16:49 local time until 9:00
Session Close: Mon Oct 31 16:50:24 2011
Received on Monday, 31 October 2011 23:57:23 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 31 October 2011 23:57:23 GMT