New PDF > Word program

Mats Wiman wrote:

I found and have downloaded a trial version of a program called Solid Converter PDF at:

http://www.solidpdf.com

It is purported to make an excellent conversion, keeping the formatting better than other programs (Acrobat 6.0?)

Please have a look and report back to this list.

Mats J C Wiman
Übersetzer/Translator/Traducteur/Traductor > swe
http://www.MatsWiman.com
http://www.Deutsch-Schwedisch.com
http://www.proz.com/pro/1749
(Proz.com moderator, deu>swe, Swedish)
Träsk 201
SE-872 97 Skog
Schweden/Sweden/Suède/Suecia
Tel:+46-612-54112 Fax:+46-612-54181 Mobile:+46-70-5769797

Kenneth Cox wrote:

The basic idea seems good (why use an OCR program when you already have text?), so I downloaded and tried the trial version on several PDF documents. SolidPDF can be used in several different modes that are intended to preserve the original formatting and layout to various degrees, but if your objective is to extract text so it can be translated, you're probably more interested in the integrity of the text than the fidelity of the layout. And with regard to this there are several features of SoftPDF that I find at least questionable, if not annoying or simply unsatisfactory.

1. It insists on converting all text in fonts not embedded in the PDF document (and/or not present on the host system, I'm not sure about this) into 'standard' fonts. Since Word will happily accept just about any font you throw at it and display it using a substitute font if necessary, it's hard to understand why this should be (considered to be) necessary (and if you copy and paste text from a PDF document to a Word document, the orignal fonts are preserved, so it is certainly possible to retain the original fonts). Replacing fonts may (severely) annoy clients who want to use the translated text in a different version of the original document and will be forced to convert it back to the original fonts, and the replacement fonts may (typically will) not have the same dimensions as the original fonts, which can cause layout problems.

2. The way SolidPDF handles tables in the original is rather unsatisfactory.
First, unless you use the mode (sorry, I forget the exact name just now)in which each block of text is placed in its own text box, it aligns texts located in the various cells of a table on a line-by-line basis using multiple spaces (nary a tab to be seen), which is naturally a nightmare for translation (maybe it doesn't do this if you select plain vanilla text extraction; I didn't try that). Of course, you can (probably) get around this by using the índividual text boxes mode, but that's not particularly convenient for translation, and it's an absolute nuisance for doing anything with the result other than printing it.
Second, SolidPDF tries to reproduce the cell boundaries (rules) of tables using Word drawing objects (lines), rather than by constructing Word tables. Besides the fact that the result will most likely be useless if you significantly edit (read: translate) the text, in many cases the extracted text does not accurately fit within the lines (probably as much as anything due to font replacement and converting all fractional font sizes to the nearest 0.5 point size).

3. Finally, at least in all the modes I tried (which did not include plain vanilla text extraction), SolidPDF tries to preserve variable word spacing in justified text by inserting multiple spaces as necessary. Of course, you can do a global find and replace to change all multiple spaces to single spaces, but that will clobber all of your pseudo-tables, bullet lists, etc. I'm far from being technically knowledgeable about PDF, but this may be a logical consequence of extracting the text info from the PDF file rather blindly (if you copy and past text from a PDF file into a Word file, you don't get the extra spaces...).