PDF Compare - Text Goes Missing

[Cobol Jim]

I work for an insurance company and we put our policy documents to PDF using a product called Gentext. When I view the PDF through Adobe the doc looks fine. The "Hardcoded" info (things designed into the form itself, like boxes, lines, headings, etc) is there, but only some of the variable data (coverages, premuims, etc) is there. I say 'some' variable data is there meaning in certain areas it is missing.

Example - Name, address, phone, coverage info, vehicle info, structure info, etc are examples of variable data. It can appear on one form but be missing from another. My only guess is where the data is missing, perhaps Gentext is using a font that BC3 doesn't recognize, therefore BC3 displays nothing? That's just a guess on my part. I don't take part in the forms creation so I don't know what the font is. What I do know is Gentext doesn't always use the true font. It migh say Courier New 10 but may or may not be accurate.

If I create the same PDF using another product called XPAF and bring that PDF into BC3 everything is there. No problems whatsoever. So, if I try to compare XPAF to Gentext it throws a difference becasue of the missing data on the Gentext copy. As I said though, view each one in Adobe and everything is there.

Any clues?

Update - the font is Palentino, 10 point, fixed pitch

[Aaron]

Hello,

If you open the PDF in Adobe, then use the File menu -> Save As Text (for either/both PDFs), what output do you see?

Beyond Compare is performing a conversion in the background to save as plain text. The font might be the issue, or it might be another aspect. We most commonly see this when the document has a picture of text, but the text is not actually select-able data. It would need to be text to be caught and converted during the save as plain text phase.

Do you have any sample files you could email in to [email protected]? If so, please include a link back to this forum thread for our reference.

[Cobol Jim]

I sent two emails with attachments.

[Aaron]

Replied by email with a few specifics. At least a couple of the trouble spots are pictures of text and not actual text data. It isn't highlightable/copy/paste, so the conversion process isn't able to interpret it into the .txt file that we compare. Please let us know (by email) if there are other specific sections of the files you would like me to investigate.