I am using Microsoft Word 2003 and Acrbobat X Pro 10.0.
My tables in Word have been set to "Heading Rows Repeat"
When I first do the check, I get this error: "Table Row is not a child of Table, THead, TBody, or TFoot."
The table is tagged as Table > TR > TH > Normal.
Then when I set the tag as "Table > TR > TH > TD", I get that error again as well as "Table Header or Table Data is not a child of TR"
Here's a screen shot of what I did: http://i46.tinypic.com/ok20ye.jpg
My document has many tables (as it is a 200 page document) and the only errors I have when I do the accessibility check is from the tables - so any help would be much appreciated.
In a Tagged PDF —
While TH and TD may both be a child to TR (cells in a table row) it is not permitted to have a TH as a child of a TD (a cell within a cell that is part of a row).
Nor is it permitted to have a TD as a child of a TH (again, a cell within a cell that is part of a row).
When either is present the error "Table Header (the TH) or Table Data (the TD) is not a child of TR" is presented via Full Checker.
From the screen capture -
I suspect the first row may be contributing to the problem
That row (containing "Record of Updates") should be a "repeat..." row selection back in Word.
In the PDF you must also set the Column Span to a value of "8".
The second row is also a header row. In the PDF remember to set each cell as a "Header cell" having a Scope of "Column".
Because the table shown has two vertical groups of "header columns" you've a problem.
If you carefully set the required Headers attribute for each table cell and the requisite ID entries for all cells a table with a single column (preferred is the first column only) holding TH cells one can get to a reasonably accessible PDF.
The screen capture shows a table with the first and the fifth columns have cells intended to serve as TH cells.
As providing proper Headers attributes and ID entries is entirely a manual operation getting the screen capture table set up correctly will be a non-trivial endeavor.
It can be done but there's still a problem.
None of the typically used AT applications (NVDA, Windows Eyes, JAWS) deal effectively with such tables.
In part this is due to ISO 32000-1 (the PDF Standard) not identifying a normative algorithm for implementers to associate headers with row and column TH cells' ID entries.
Fortunately this will be addressed in ISO 32000-2.
Having a copy of ISO 32000-1 on hand can be most helpful.
An ISO authorized copy is available from Adobe.
Adobe's ISO 32000-1
Section 14..18.104.22.168 Table Elements and Section 14..5.7 Table Attributes are key to understanding Tagged PDF's Table element/tag and the allowed child elements/tags..
Back to the possible source of the issue(s).
I'm thinking something is awry back in the Word authoring file.
I'd have to have a file for a look-see to be sure.
Something I'd do. Get "Record of Updates" out of the first row; dump that row.
Have "Record of Updates" as, perhaps a Word Heading 3 ("Section 1 ..." as a Word Heading 1 and "Overview" as a Word Heading 2).
Of course one would use Word's built-in Headings as with PDFMaker's tag management assures proper role mapping of these to PDF's Heading elements/tags.
The idea is to have "Record of Updates" be a table caption.
Unfortunately Word does not appear to have a built-in "table caption/title" still that could be programmatically role mapped to the "Caption" element/tag.
Just have the No. , Date, ... header row.
Have only one "No." column.
First row is your "header row" and all cells are "TH".
All subsequent rows are "data rows" and all cells are "TD".
Thanks! I did what you said at the end of the post and I didn't get any error messages when I did the check.
Here's how I made it: http://i46.tinypic.com/34qv12s.jpg
But that was just one table out of 190 pages, so hopefully the tables work out fine - prob post back here again.
Anyway, thanks again.
I have another question:
When I try to convert my document (all 200 pages), I get a message saying "insufficient memory" - which I know is becasue of my computer ... but it still tags some of the document. And then when I do the accessibility check, I get those same errors on the table even when following the steps.
When I just convert that table (see screen shots above), I don't get that error and when I do the check, everything works properly.
So, I'm assuming it is because Adobe cannot convert the document fully and properly, so that's why the errors when I do the check still come up even when I do all the steps in Word to correct it?