Tagging Moves Text in Tables

Question

I'm in the process of converting hundreds of PDFs to be 508 compliant and one of those steps includes adding tags to the document.  What I'm noticing is that the tagging is moving some of the text to the bottom of or outisde of some of the tables.  Does anyone know what is causing this or how I can prevent it?
Thanks!

Answer 1

The Add Tags to Document follows general rules for identification of an untagged PDF's logical hierarchy.
For Tables, this presumes that, in the authoring file, the authoring application had an "insert table" feature and that this feature was used.
If (as was done rather often "back in the day") a table was simulated by use of space and/or tab characters to position what was to represent a "table" then the PDF does not contain a "table" per se.
If a "table" was inserted the cell input may have been positioned with space/tab characters rather than using the application's "style" configuration (for indents, white space & such).
For simple tabular/data tables the Add Tags to Document does the job (even if the table in the authoring file was a "layout" table (no defined header row)).
However, complex tables won't "fit" the general rules.
Complex having something like:
Multiple "header" rows (either vertical or horizontal or some of both)
Spanned cells
Aside from having to deal with what came from those old authoring files remember to add a Table Summary for each Table and for each header cell <TH> its requisite value for Scope.
As well, for spanned cells the appropriate value for row or column Span.
For "complex" tables you get into assigning the Table Attribute of "Header" and associate "ID" entries.  A neat exercise; but, better avoided when creating content by breaking tables out to simple tables rather than multiple tables mushed together into one.
As you are working on "legacy" PDF the only spot on method of bringing the content to workable Accessibility may well be manual tagging as a "best guess" from Add Tags to Document has limitations (e.g., It cannot provide Alternate Text for content in a <Figure> element.).
So, when climbing about in the structure tree (you'll have to go there to get it done) remember to save as and increment the file name (_01, _02, or some such) as there is no "undo".
You've a "doable" task — it just takes time, a good chunk of time as you've a large population of PDFs.
The PDF(s) for rosebird took somewhere between 70 to 100 hours.
http://acrobatusers.com/forum/how-make-pdf-ramseyer-format-accessible
So, you may want to address the level of effort/time reality.
Of course, for $$ the you could out source the task to one of the several companies that do this sort of thing (would not be "me" - I'm not affliated with any of them ).
Be well...
Message was edited by: CtDave