This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.

Bug 5561 - SML should define a Simple XLink Reference Scheme
Summary: SML should define a Simple XLink Reference Scheme
Alias: None
Product: SML
Classification: Unclassified
Component: Core (show other bugs)
Version: CR
Hardware: PC Windows NT
: P2 normal
Target Milestone: CR
Assignee: Len Charest
QA Contact: SML Working Group discussion list
Depends on:
Reported: 2008-03-12 19:34 UTC by Pratul Dublish
Modified: 2009-06-08 16:53 UTC (History)
5 users (show)

See Also:

Len's proposed changes to the XLink ref scheme note (29.97 KB, application/vnd.openxmlformats-officedocument.wordprocessingml.document)
2009-03-06 07:51 UTC, Len Charest
Len's proposed changes to the XLink ref scheme note (PDF) (279.32 KB, application/pdf)
2009-03-06 07:52 UTC, Len Charest
docx (32.73 KB, application/vnd.openxmlformats-officedocument.wordprocessingml.document)
2009-03-20 14:15 UTC, John Arwe
pdf (116.40 KB, application/x-download)
2009-03-20 14:16 UTC, John Arwe

Comment 1 Pratul Dublish 2008-03-17 14:40:08 UTC
Imho, the SML WG is not in the business of defining reference schemes. The SML spec is completely independent of any specific reference scheme (including the sml:uri scheme). It defines (in Section 4.3) a set of requirements that an SML reference scheme must comply with, and other W3C  WGs or community members can define a new SML reference scheme for their domain/scenarios by complying with these requirements. The SML WG defined the sml:uri scheme because of two reasons:

1.	It is the basis for interop using SML IF  
2.	The SML WG felt that it will be the most commonly used scheme  
Comment 2 Pratul Dublish 2008-04-02 20:38:28 UTC
Resolution in 4/2 F2F meeting: the WG will create and publish a note on SML XLink Reference Scheme.  
Comment 3 Kumar Pandit 2008-04-11 03:54:00 UTC
Here is link to the email sent by Pratul that has a sample xlink based scheme.

Note that the document sent by Pratul is a work in progress. It has not been reviewed and approved by the SML WG members.
Comment 4 Henry S. Thompson 2008-04-18 13:03:47 UTC
With respect, we already knew that anyone _could_ define a XLink reference scheme.  The point was that if the WG doesn't include it in the spec., it won't get used because it won't be interoperable.  I'm sorry the WG doesn't see it as an important selling point for SML, but I won't lie down in the road over this.
Comment 5 Pratul Dublish 2008-04-24 19:26:20 UTC
The WG appreciates your flexibility on this issue and expects that the proposed note will facilitate interoperability of implementations of the XLink reference scheme

Comment 6 Kumar Pandit 2008-05-01 19:15:51 UTC
resolution in conf call (5/1/08): add externalComment and remove decided from keywords
Comment 7 Pratul Dublish 2008-10-29 18:48:51 UTC
Draft note sent to public list
Comment 8 Len Charest 2009-01-15 19:42:10 UTC
Note has been converted to HTML and put under version control in CVS. See
Comment 9 Len Charest 2009-03-06 07:51:18 UTC
Created attachment 657 [details]
Len's proposed changes to the XLink ref scheme note

Open in Word 2007.

Change tracking is on in the Word doc, so insertions and deletions are visible.
Deletions are marked in red with a strike-through line. Insertions are
double-underlined. Where I have replaced text, the insertion always follows the
deletion and there is an explanatory comment attached to the insertion.
Comment 10 Len Charest 2009-03-06 07:52:47 UTC
Created attachment 658 [details]
Len's proposed changes to the XLink ref scheme note (PDF)

Open with a PDF reader.
Comment 11 John Arwe 2009-03-20 14:15:55 UTC
Created attachment 667 [details]

In reply to comment #9
Comment 12 John Arwe 2009-03-20 14:16:19 UTC
Created attachment 668 [details]

In reply to comment #10
Comment 13 Virginia Smith 2009-03-20 22:32:01 UTC
The following are my comments on the [style] comments - referencing original comments from attachment in Comment #10 of Bug 5561:

title - name should match SML spec which would be "SML XLink Reference Scheme".

link style comments - In the SML and SML-IF specs, the convention is 'generally' (meaning I don't think we did a thorough check to make sure we were strict about this) to use [] when the link is parenthetical to the sentence and not use [] when the text is part of the sentence. There are exceptions (such as for Schema elements). However, I think that we should maintain this convention.

Definition section - the subsections should match SML spec 4.3 as much as possible.

Notes - the SML and SML-IF spec show the 'Note' as separate from the actual note content. This is governed by a W3C stylesheet and I prefer to see consistency with the specs. In other words, the "Note" is separate from the paragraph as originally written.

LKC12 - I disagree (agree with John's comment)
LKC17 - I disagree and prefer 'if'
LKC19 - I prefer to keep the text.
LKC25 - I recommend just adding 'valid' to paragraph 3.0.
Comment 14 Kirk Wilson 2009-03-22 14:41:48 UTC
The following are my comments - referencing the original comments.

LKC1: Title should follow the style SML xxx Reference Style to be consistent with the SML spec.

LKC7: The full title of section 2 should be restored.  Simply entitling the section "Definition" and depending on the context for what the definition is of seems to be an inconvience to the reader.

LKC8: I would vote against restructing section 2.  Since it is relatively simple to define an SML XLink reference scheme (as opposed, e.g., to an EPR-based SML reference scheme), we should keep the structure as close as possible to that of the SML reference scheme definition.

LKC14: The "active voice" modification is actually not "active voice", but an imperative: "Obtain a document. . ".  This seems to be taking a step into the "algorithmic" form of expressing the defintion, which the WG has rejected--with VERY good reasons (IMHO).

I agree with John that LKC19 (D has no value) is not irrelevant and LKC22 is correct as it stands ("or" vs. "and").  Good catch on the sml:targetType.
Comment 15 Len Charest 2009-04-06 08:18:32 UTC
Revision 1.4 has been checked in to CVS. See

Note that the references section is still TBD.
Comment 16 John Arwe 2009-04-06 14:03:03 UTC
2.1.a singular/plural disagreement
> two attribute information items whose local name are type and href, and whose namespace name  is 
- two...items...local name are..., ...ns name is...
should be "local nameS" are, ns nameS are

Otherwise the changes look ok to me.
Comment 17 Len Charest 2009-04-10 00:22:27 UTC
(In reply to comment #16)
> 2.1.a singular/plural disagreement
> > two attribute information items whose local name are type and href, and whose namespace name  is 
> - two...items...local name are..., ...ns name is...
> should be "local nameS" are, ns nameS are
> Otherwise the changes look ok to me.

I understand the problem, but I'm not sure of the solution. The terms "local name" and "namespace name" are surrounded with square brackets (since they are abstract properties). Does the "s" go inside the brackets or outside?

IOW, should we write:
    "It has two attribute information items whose [local names] are type and href, and whose [namespace names] are"

    "It has two attribute information items whose [local name]s are type and href, and whose [namespace name]s are"

Or perhaps restructure the sentence to avoid this issue.

Comment 18 Virginia Smith 2009-04-13 16:34:47 UTC
Looks fine to me with the following change:

Table of Contents:
2. SML XLink Reference Scheme Definition
3. Examples

Remove the "Â ".
Comment 19 Kirk Wilson 2009-05-11 17:00:09 UTC
1.  The first sentence of the Abstract should specify references to the XML and XML Schema specifications.  Bibliographical references to these two specifications should be added to the Reference section.
2.  The first sentence of the Introduction should include mention of XML (along with XML Schema, i.e., the sentence should parallel the first sentence of the Abstract), and specify references for both the XML and XML Schema specs.
3.  The URI to the list of patent disclosures in the Status section is incorrect.  It should be
Comment 20 Len Charest 2009-05-22 16:49:52 UTC
I've made changes in response to Comment #16 and Comment #19. See revision 1.7 in CVS ( 

The Status section has also been modified to match the text of the EPR note (as appropriate).

The References section has also been reworked.
Comment 21 Len Charest 2009-05-22 22:58:27 UTC
Revision 1.8 checked in to CVS. Misc tweaks to mark-up:

  - Normalized names of same-document reference targets to ref-XXX
  - Wrapped all references with <cite/> (rather than <i/>)
  - Restored Back to Contents jump links on section headings
  - Fixed broken links
Comment 22 Kirk Wilson 2009-05-31 20:10:38 UTC
Minor comment:

In section 2, items 3,d,ii,1 & 2, identifying the target T is said to depend on a "XML-Schema determined ID".  The XPointer spec refers to this simply as a "schema determined ID".  Given the definition of a schema determined ID, the specific reference to "XML Schema" is unnecessary.

Suggest a simple editorial change to "schema determined ID" (with a small "s") to remain consistent with the XPointer spec.
Comment 23 John Arwe 2009-06-01 15:36:17 UTC
(In reply to comment #22)


> The XPointer spec refers to this simply as a "schema determined ID".  

IIRC from the SML barename discussions (see, I do listen occasionally!), that term includes not only XML Schema but also DTD-defined IDs.  Since all XML parsers are required to process the internal DTD subset, allowing DTD-determined IDs imposes no new processing requirements.  Saying "XML Schema determined..." would have a different (narrower, and therefore incorrect) meaning.
Comment 24 Len Charest 2009-06-01 18:02:35 UTC
I've checked in revision 1.9 with the following editorial changes:

  - Changed publication date to 11 June 2009 and updated URIs to match
  - Section 2, item 3.d.ii: Replaced phrase "XML-Schema-determined" with
  - Section 2, item 4: Replaced phrase "the applicable URI RFC" with "the
    applicable RFC for URI generic syntax"
  - Updated entry for [XML] in References section
  - Moved entry for [XPointer] so that entries are in alphabetical order
Comment 25 John Arwe 2009-06-08 16:53:52 UTC
Per the telecon of June 1, the working group has decided it has finished its work on this Note.

Per the telecon of June 8, I am closing out this bug.