Boeing Proposal for WebCGM 2.0 Fixes before LC
28 May 2006
Submitted by: David Cruikshank (Boeing)

This document identifies a collection of errata, clarifications, and corrections to the W3C-published Submission version of the WebCGM 2.0 specification. Boeing believes that these changes will correct known errors, and will improve the clarity and accuracy of the specification. Boeing proposes that they be applied by the W3C WebCGM WG before going to Last Call review, and believes that doing so will facilitate the review and reduce adverse comments.

Format: Proposed change, brief rationale, references. Each change points to an illustration of the change, implemented in an edited version of the Submission, and the location of the change highlighted like this. The edited version of the Submission was prepared by the Chair of the WebCGM WG.


Proposed change 01:
Add a new toNodeList() method to the existing WebCGMAppStructure interface. In the DOM chapter (ch.5) modify the IDL for WebCGMAppStructure and add the method description as proposed, and apply the corresponding change to the ECMAScript binding.
Rationale:
This new convenience method is the least disruptive solution to the identified problem, that the highlight() method doesn't allow easy highlighting of a single node.
References:

Proposed change 02:
Fix typos in existing WebCGMPicture interface description section: getAppStructureById returns WebCGMNode, not boolean, and add missing description of state parameter of highlight().
Rationale:
Editorial errors.
References:

Proposed change 03:
Add the usual extensibility entities to the 'linkuri' element of the existing XCF definition (4.3.8), to the complete DTD of section 4.4, and to the standalone DTD (webcgm20.dtd) that is posted with the document.
Rationale:
Extensibility of 'linkuri' is a requirement of S1000D that was overlooked in the preparation of the OASIS CS and W3C Submission text.
References:

Proposed change 04:
Fix typo (missing quotes) to existing fragment EBNF.
Rationale:
Editorial errors.
References:

Proposed change 05:
Fix typo ("v2.0") in DOCTYPE example of existing XCF namespace (4.2.3) section; remove deprecated value "2" from valid compression types list of TILE (T.19.29) in existing PPF.
Rationale:
Editorial errors.
References:

Proposed change 06:
Change existing mapping specification for the deprecated old object behaviors 'highlight' and 'highlight_all' to 'full+newHighlight'.
Rationale:
'full+newHighlight' corresponds exactly to the old highlighting behavior. 'newHighlight' alone leaves the navigation behavior to be determined by the picture-load operation, which should yield the same result, but the exact-match object behavior 'full+newHighlight' puts is a more precise all-in-one-place mapping.
References:

Proposed change 07:
Change the existing declared type of getAppStructureById to WebCGMAppStructure in the IDL and description (5.7.5), and the ECMAScript (ch.8). (And also fix the cut-paste error in the description, about boolean return value.)
Rationale:
If the return type is WebCGMNode instead of WebCGMAppStructure, then none of the attributes and methods of the WebCGMAppStructure interface can be applied to the returned value (which is an application structure), and the given examples like in 5.7.1 and 5.7.6 don't work.
References:

Proposed change 08:
Clarify the existing 'src' attribute description (on WebCGMMetafile interface), especially regarding rules against picture behavior in this context.
Rationale:
The wording of the paragraph is unclear and confusing.
References:

Proposed change 09:
  1. restore the WebCGM 1.0 pictseqno ::= (digit)+ production in the existing fragment EBNF (3.1.1.2),
  2. restore the WebCGM 1.0 out-of-range specification for pictseqno in the existing picture selection description (3.1.2.1), and rewrite the introductory paragraph of 3.1.2.1 to explain about mixed-version 1.0/2.0 environments.
  3. fix existing conceptual description (2.3.6) about WebCGM 2.0/1.0 picture linking, and clean up occurrences of similar wording in Ch.3.
Rationale:
The limiting picseqno to 1 in the fragment syntax was a side effect (unintended) of limiting metafile content to one picture in WebCGM 2.0 metafiles. Since fragments can originate in HTML content pointing into metafiles (of unknown version), and 2.0 viewers otherwise need to function in legacy or mixed 1.0/2.0 environments, WebCGM 2.0 needs realistic rules to accomodate legacy and mixed situations.
References:

Proposed change 11:
Add UTF-16 (RFC 2781) reference to existing normative references (1.2).
Rationale:
Support of unicode UTF-16 in graphical and non-graphical text is a normative requirement of WebCGM.
References:

Proposed change 12:
Rewrite existing XCF-relative-URI description (3.1.1.5) to clarify that this is only about the resolution of an XCF reference within the fragment of a WebCGM reference.
Rationale:
That distinction was not clear enough, and the section was taken to be at odds with RFC-3986 p.28.
References:

Proposed change 13:
Revise existing XCF-application description (3.1.2.6) to remove apparent contradiction and clarify that the described actions are required only of WebCGM DOM-supporting viewers.
Rationale:
As written, two sentences appear to contradict each other.
References:

Proposed change 14:
Replace the table in the existing "Summary of behaviors" (3.1.2.7) with informative prose that clarifies the purpose and references of the section.
Rationale:
The table has been widely criticized as being incomprehensible, and misinterpreted as introducing normative requirements, whereas it is merely intended as an informative table of contents and directory to the widely scattered normative rules.
References:

Proposed change 15:
Update the PNG reference in existing "Normative references" (1.2).
Rationale:
It is out of date and incorrect.
References:

Proposed change 16:
Add further discussion to existing drawing model description of CGM/WebCGM (2.2.2).
Rationale:
The drawing model description in the WebCGM spec is not very detailed, and one must search in the ISO CGM standard and the ISO Register of Graphical Items to piece together the details.
References:

Proposed change 17:
Fix editorial errors in existing "Overlaying a picture" section (2.2.3).
Rationale:
"Object tag" should be "object element", and the param element is "background", not "TRANSPARENT".
References:

Proposed change 18:
Make changes to existing para/subpara conceptual description, to content conceptual description, to normative 'para' & 'subpara' descriptions, and to normative 'content' description, to clarify the data model and intended usage of the elements, their relationship to (application-defined) text search, and to more clearly separate normative from recommended (or informative) aspects.
Rationale:
The (WebCGM 1.0) data model corresponding to para/subpara/content has proved confusing to several WebCGM reviewers, and there is further confusion about providing support features for application defined text search functionality versus normative capabilities and requirements of WebCGM itself (as defined in WebCGM 1.0).
References:

Proposed change 19:
Fix the first paragraph of the existing hyperlinking conceptual description (2.3.6) to remove "bi-directional" implication and clarify what was actually meant.
Rationale:
Bi-directional has implications in linking technology that were not intended in this informative description (WebCGM 1.0 text).
References:

Proposed change 20:
Revise the 3rd paragraph of the existing hyperlinking conceptual description (2.3.6) to clarify about rfc3986 and point off to appropriate normative sections.
Rationale:
This (informative) paragraph was overlooked when changes were made to the normative parts (3.2.2.3 and subsections of 3.1.1).
References:

Proposed change 21:
In the two executable examples of existing WebCGMEventListener (5.7.9) and WebCGMEvent (5.7.10), change the 5th line to "var cgmDoc;". Make same changes in ex_WebCGM_Event.html and ex_WebCGM_EventListener.html
Rationale:
Incorrect ECMAScript.
References:

Proposed change 22:
Add Unicode to existing normative references section (1.2), in both "generic" and "specific" form, add ISO 10646 reference in "generic" form, and add current UTF-8 & UTF-16 references.
Rationale:
WebCGM has normative dependence of these.
References:

Proposed change 23:
Change listing of existing Example of 5.7.5 so that mime type is text/ecmascript.
Rationale:
IE browser does not support application/ecmascript, although it is preferred by IANA (see "issue pointer").
References:

Proposed change 24:
[No specific-section proposal or sample implementation yet.] Change "URI" to "IRI" where appropriate; consider adding some explanation similar to Tiny 1.2.
Rationale:
An informal comment from CL, that has not yet been fully processed.
References:

Proposed change 25:
[No specific proposal or sample implementation yet.]
Rationale:
An informal comment from CL, that has not yet been fully processed.
References:

Authors: David Cruikshank, Boeing
Comments to: David Cruikshank
Last Updated: $Date: 2006/05/28 21:37:15 $