|
This represents a change from the last published draft text.
I think it more nearly captures the meaning of this attribute as setting the boundary for the collection of index entries to either a chapter or the entire document. I am setting this issue to resolved but welcome further discussion on how to make our meaning clear. Since the bafflement in the defect report is about what is meant by "current chapter," I doubt that changing the text to speak of "a chapter" is edifying.
I think the explanation in Patrick's comment is better than in the proposed resolution. However 4.1.1 does not define what heading is for a chapter, although one might infer it is the one with text:outline-level="1". This is actually left to be inferred. Is it to be understood that the chapter consists of the material between the nearest preciding heading with text:outline-level="1", if any, and the nearest following one with text:outline-level="1", if any, in the vicinity of the <text:table-of-content> element? Is a preceding one with text:outline-level="1" included in the TOC or are only internal heading levels covered in the chapter-level index? Can even narrower chapter index scopes (that is, beneath an arbitrarily-deep subchapter) be selected depending on what the text:outline-level" value, if any, is on the nearest preceding <text:h> element? I think the "current chapter" question needs to be answered directly for the defect report, if it can be, and then tidied up in ODF 1.2 as appropriate. Another solution would be to solve it in ODF 1.2 first and then retrofit if still needed into, say, the aligned 1.1. I notice that this problem comes up again in the first sentence of (ODF 1.0) 7.4, 7.4.1 (Index Scope again), 7.5, 7.6, 7.6.1 (Index Scope again), 7.7.1 Index Scope again, cross-referenced to 7.4.1), 7.8.1 (see 7.4.1 again).
I also notice that chapter-number is often at any level of sub-chapter in its use throughout the specification. That has me ask again about how low one can scope a table of contents in a deeply-nested sub-chapter. For example, 1.2.3 is given as an example of a third-level chapter number. Can these index-scope schemes work at any chapter level? I also see ambiguity with regard to text:start-numbering-at for Numbering Schemes which can apparently run at a chapter level for footnotes, for example. I'm afraid to look at ODF 1.2 Part 1 CD04 by searching on "chapter" the way I just did in ODF 1.0. Issue
Issue We might want to say, in the Errata 01 CD05, that the behavior when the optional text:index-scope attribute is omitted is implementation-dependent. Alternatively, we might want to punt the whole issue to future work (i.e., ODF 1.2 and Answering Dennis question, the chapter is being defined by 4.4.1 Headings
"Headings define the chapter structure for a document. A chapter or subchapter begins with a heading and extends to the next heading at the same or higher level." Dennis, if you think you might have possibly noticed additional problems during review, please address them for the latest ODF specification (or ODF Next). The errata document covers the issues of the national body and is not meant to extend them. Changed resolution due to comment of
http://tools.oasis-open.org/issues/browse/OFFICE-2734 Here is the problem with this.
First, in ODF 1.2, the following statement is proposed. Wording like "When the value of the attribute is "chapter" the index is constructed for the chapter that contains the index. A chapter consists of all those paragraphs and headings which have the same immediately-preceding heading with outline level 1." This is from the generic definition of text:index-scope in section 19.814 of ODF 1.2 CD05 Part 1. There are two difficulties with this. First, we really mean "the table of contents is constructed for the chapter that contains the <text:table-of-content> element." The second difficulty is that the <text:table-of-content> element can occur in some very unusual places for which it is not exactly clear how this is intended to work. Here are the places that are strange: Inside a <text:index-body> that occurs in a <text:table-of-content> element and in a variety of other index elements, wherever those indexes are. In a <draw:text-box>, in a <table:table-cell>, in a <text:note-body>, and in <style:header>, <style:header-left>, <style:footer>, and <style:footer-left>. There is no particular problem with occurrences in a <text:deletion>, in <office:text>, and in a <text:section> except <text:section> might appear in some unusual places as well. Because of these difficulties, I would make the following new statement in the Errata for ODF 1.0 section 7.3.1:Index Scope: "The text:index-scope attribute specifies whether index entries from an entire document or from the chapter that contains the <text:table-of-content> element are used to construct a table of contents. The default and how the chapter is determined for the different ways a <text:table-of-content> element may occur is implementation-dependent." Dennis and I agreed on a new resolution.
|
|||||||||||||||||||||||||||||||||||||||||
 |
|
Bug
Applied
Major
ISO/IEC JTC 1/SC 34 N 1078 : DEFECT REPORT NUMBER JP2-5
Current draft text reads: "The text:index-scope attribute specifies whether the table of contents is generated for the whole document, or only for a chapter."
Note, not the "current" chapter.
Understand that "chapters" are defined by the presence of <text:h> elements. (In ISO 26300, see 4.1.1, current draft, see 4.1.1).
The setting "chapter" means that only the material in the chapter where the <text:table-of-content> element occurs will be used in the construction of that particular table of contents. A setting of document means that the table of contents for the entire document will be constructed.
I have a proposal with a comment.