RE: [cgmo-webcgm] QUESTION: ACI DTD and conformance

[Lofton Henderson <lofton@rockynet.com>]

Rob --

Good.

I was of course aware of the extensibility wrinkle as I compared ch.4 and 
ch.9, and the ACI is not extensible like the XCF.  But I didn't put 
two-and-two together and see that as the reason for the *if* in the Ch.4 
conformance statement.

-Lofton.


At 01:50 PM 4/22/2009 -0600, Robert Orosz wrote:
>[...]
>XCF has the wrinkle (not present in ACI) that it is extensible by the user.
>Stub parameter entity references are declared in the XCF DTD which users
>could define in their own DTDs.  I believe that in order to get this to work
>properly, you would need a document type declaration that referenced your
>own DTD first so that the appropriate parameter entities are defined, and
>then you reference the WebCGM DTD in order to complete the document type
>declaration.  Hence you have the more complicated conformance statement in
>4.2.4, but it would not be needed in the ACI case.
>
>Rob
>
>-----Original Message-----
>From: Lofton Henderson [mailto:lofton@rockynet.com]
>Sent: Wednesday, April 22, 2009 1:10 PM
>To: CGM Open WebCGM TC
>Cc: David Cruikshank
>Subject: RE: [cgmo-webcgm] QUESTION: ACI DTD and conformance
>
>
>Thanks, Rob, for the response and analysis.
>
>I don't see any need or use for public identifier.  Does anyone else?
>
>Accordingly, we would add to Ch.9, similar to 4.2.3, a "System Identifier
>for the WebCGM 2.1 ACI" and "DOCTYPE Example":
><!DOCTYPE webcgmConfig SYSTEM
>"http://docs.oasis-open.org/webcgm/v2.1/webcgmConfig21.dtd";>
>
>And make it part of a validity requirement in 9.2.2.  (That DTD location
>has precedent in WebCGM 2.0 for the XCF DTD.)
>
>One question about #4 (below):  why does 4.2.4 say "...if an appropriate
>document type declaration ... which points to the WebCGM DTD is included
>immediately after the XML declaration ... the result is a valid XML
>document."
>
>I know this language was pinched from similar language in SVG, but can't
>recall the rationale behind the "if".  For ACI at least, do we want
>conforming ACI instances to contain a DOCTYPE or not?
>
>-Lofton.
>
>At 11:15 AM 4/22/2009 -0600, Robert Orosz wrote:
> >Lofton,
> >
> >I agree with your observation that [1] below is under-specified.  We
>specify
> >a DTD in section 9.4, but nowhere do we state that an ACI file must be
>valid
> >(conforms to the DTD).  Here are my answers to your questions in order.
> >
> >1) Yes.
> >2) This depends on whether we want to specify a public ID for ACI files.
>In
> >my opinion, a public ID is not really necessary, but we might want to
> >specify one anyway since we did so for XCF.  If we had a public ID for ACI,
> >the DOCTYPE would be similar to the XCF DOCTYPE.  If not, it would look
> >something like this.
> >
> ><!DOCTYPE webcgmConfig SYSTEM
> >"http://docs.oasis-open.org/webcgm/v2.1/webcgmConfig21.dtd";>
> >
> >3) Yes.  As I recall, OASIS has established rules for where DTDs are placed
> >in their public web space.  We dealt with this during 2.0 spec development
> >when XCF was introduced.
> >4) I think it is sufficient to say that the ACI document must be valid.
> >This is a well defined term in the XML specification.
> >
> >BTW, the first bullet in [3] below mentions XML 1.0 3rd Edition.  XML 1.0
>is
> >now at 5th Edition.
> >
> >Rob
> >
> >-----Original Message-----
> >From: Lofton Henderson [mailto:lofton@rockynet.com]
> >Sent: Wednesday, April 22, 2009 10:36 AM
> >To: CGM Open WebCGM TC
> >Cc: David Cruikshank
> >Subject: [cgmo-webcgm] QUESTION: ACI DTD and conformance
> >
> >
> >All (esp. our markup experts Stuart, Rob, Dave) --
> >
> >While looking at Don's ACI files, I noticed this:
> ><!DOCTYPE webcgmConfig SYSTEM "webConfig.dtd">
> >
> >Small problem: there seems to be a naming error.  In the current-editor
> >directory, the file name is "webcgmConfig21.dtd".  (Don, please don't
> >update this on the FTP site yet, until we sort through the rest of the
> >issue.)
> >
> >Bigger QUESTION:  what should such a DOCTYPE look like and what about the
> >conformance statements in Ch.9?
> >
> >References:
> >
> >ACI DOCTYPE and conformance:
> >[1]
> >http://www.w3.org/Graphics/WebCGM/drafts/current-editor-21/WebCGM21-Config.
>h
> >tml#ACI-conform
> >XCF DOCTYPE:
> >[2]
> >http://www.w3.org/Graphics/WebCGM/drafts/current-editor-21/WebCGM21-XCF.htm
>l
> >#namespace
> >XCF Conformance:
> >[3]
> >http://www.w3.org/Graphics/WebCGM/drafts/current-editor-21/WebCGM21-XCF.htm
>l
> >#conforming
> >
> >Discussion:
> >
> >First, compare [2] & [3] with [1].  It seems to me that [1] is a bit
> >under-specified.
> >
> >1.) Should [1] show what a proper ACI DOCTYPE looks like, similar to [2]?
> >2.) If "yes", then ... what does it look like?
> >3.) Further to #2, shouldn't ACI DTD references be able to point to a
> >*permanent* central location similar to [2] for XCF, i.e., something like:
> >http://docs.oasis-open.org/webcgm/v2.1/webcgmConfig21.dtd
> >4.) Should 9.2.2 [1] have a conformance bullet similar to the 3rd bullet of
> >[3], i.e., something like:
> >"if an appropriate document type declaration (i.e., <!DOCTYPE webcgmConfig
> >... >) which points to the webcgmConfig21 DTD is included immediately after
> >the XML declaration (i.e., <?xml...?>), the result is a valid XML
>document."
> >
> >Thoughts?
> >
> >Opinion:  I don't see any big issue here, but it appears that these bits of
> >Ch.9 are somewhat "rough draft" form, and can be tightened up now in
> >anticipation of this eventually becoming a public standard.
> >
> >-Lofton.
> >
> >
> >---------------------------------------------------------------------
> >To unsubscribe from this mail list, you must leave the OASIS TC that
> >generates this mail.  Follow this link to all your TCs in OASIS at:
> >https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php
> >
> >---------------------------------------------------------------------
> >To unsubscribe from this mail list, you must leave the OASIS TC that
> >generates this mail.  Follow this link to all your TCs in OASIS at:
> >https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php
>
>
>---------------------------------------------------------------------
>To unsubscribe from this mail list, you must leave the OASIS TC that
>generates this mail.  Follow this link to all your TCs in OASIS at:
>https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php
>
>---------------------------------------------------------------------
>To unsubscribe from this mail list, you must leave the OASIS TC that
>generates this mail.  Follow this link to all your TCs in OASIS at:
>https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php