Table 33-A caption contains a note to editors: "(MANUALLY SET TABLE NUMBER TO PRESERVE TABLE NUMBERING)"

[aaronmbauman]

In Section 13.6.1.1, the table of dfdl:textNumberPattern BNF syntax has the caption:

Table 33-A dfdl:textNumberPattern BNF syntax (MANUALLY SET TABLE NUMBER TO PRESERVE TABLE NUMBERING)

The parenthesised text is an instruction to whoever edits the document, not part of the specification, and it is printed in the published PDF.

In GFD.240 this caption is not a caption at all: it is the sixteenth row of the table itself, which is how the number "Table 33-A" came to be typed by hand rather than generated. The AsciiDoc conversion turns it into a real auto-numbered caption, so the table now numbers itself correctly, but the editorial note came along with the title text.

Two things to decide:

1. The note should come out of the caption.
2. Whether the table should keep a distinct number. It exists so that the tables after it keep their original numbers; with generated numbering that is handled automatically, so "33-A" may no longer be needed.

This also blocks a formatting fix. The notation lines in that table are indented in GFD.240 with the descriptions aligned on a tab stop. Restoring the alignment means making them a nested table or a definition list, and either breaks fidelity-check, because the checker reads the source's caption row as part of the cell text. Only the indentation could be restored.

[mbeckerle]

Clearly we should take the spurious text out.

And if the automatic numbering will assign the correct table numbers in the asciidoc flavor that match the MS-Word prior version, then that's clearly good enough for now.

However,.... longer run what to do about the table numbering, automatic table numbering from 1 at the beginning of the document, is not terribly scalable. Numbering by major section i.e., Table 10-1, 10-2, etc. would be more scalable and would make it less important to preserve exact table numbers within a section if you needed to insert a table as it wouldn't change every other section.

Section 11 (Simple Types) is the big problem here. It's a giant section with tons of tables having table numbers, and it's a likely place where a new table might get added.

Consider that maybe we just don't need table numbers. They can inherit the section number that surrounds them. I mean we don't have a lot of sections that have multiple tables in them do we? Typically a table involves sufficient discussion that it has a section with prose around it as well. So "The table in section X.Y.Z" may be as good, or better than, a table number.

Figure numbers have the same issue, we just don't have many figures so it doesn't come up.

[aaronmbauman]

Happy to take it out, but I haven't been through an ISO submission and you have, so I'd rather get your read first.

It is published text, page 130 of GFD.240 and so in 23415:2024 as well. But it doesn't say anything. It's an instruction to whoever edits the document that leaked into print, so removing it changes nothing normative.

That reads to me like the #88 case, where you said a minor typo needs no changelog mention or change tracking. Does this one land the same way?

One part of it is ours rather than the original's. In GFD.240 that text is the sixteenth row of the table, a hand-typed way to force "Table 33-A". The conversion makes it a real caption, so those words now sit in a caption where the original had a row.

Take it out now, or hold it for errata with a changelog entry? Either way I'll follow the same pattern next time.

[mbeckerle]

We can just change it later, but this is typographical correction so does not need a change log entry.

I don't know that this is in any of the actually published GFD.240 nor the published ISO spec. The MS-Word you started from was a new draft update in preparation, not something actually published yet.