Does anyone know of a Linux program that can be invoked from the command line to convert a PDF file to MS Word format? The PDF files are almost entirely text. The resulting files must be editable using MS Word.

pdftoword is not free but is an option. There's also pdftotext (may already be installed; it's part of the poppler-utils and xpdf-utils debian packages that ubuntu likes) which generates text files (UNIX style; no RETURN on end of line) that can be opened by msword.

Mike, thanks for the info, but I'm confused. When I google "pdftoword" I find a Windows program at www.verypdf.com. Is there a linux version of this, too? Could you send me a link to the Linux download?

The other program, "pdftotext" is probably not suitable for our needs since we want the resulting Word doc to look as much like the original PDF file as possible, including a graphic logo and signatures.

Sorry, Steven, I assumed you just wanted the text. There's not much available for Word under Linux but then think of the mentality: converting a proprietary format to another proprietary format just isn't kosher or The GNU/Linux Way or something (is this religious or political?)

Any way, I couldn't find anything under Linux except for a lot of other people looking to do the same thing. There may be a workaround though: there is a PDF to HTML converter available at http://pdftohtml.sourceforge.net/ that will get you a document editable by Word or other (Linux) word processors or HTML editors. The above page shows an example of a PDF document converted to HTML.

I'm looking at html2doc which is actually a php script to get the file all the way to Word format. Oddly enough, it still requires Windows but there seems to have been a linux conversion that I'm tracking down now. However, things just got real busy here and I might not finish before I leave for the weekend.

Thank you for doing all that research, Mike. I've concluded that there is no good way to use Linux to convert a PDF to Word format. Instead, I will have to use one of the many Windows-based programs and interface with it.

Essentially, one A-Shell program will place PDF files on a Samba share where a Perl program will detect and submit them to the converter. While another A-Shell program polls a results location to retrieve and process the DOC files. (Can an A-Shell program do what the Perl program would do? That is, invoke a Windows program running on a network server?)

It depends on the mechanism/interface required. You said that the Perl program will "submit them to the converter". If the converter is an executable running on the local machine, no problem., just HOSTEX it. If the submit is itself a file-based operation, again, no problem. If it is a TCP or FIFO, you can use TCPX.SBR or FIFO.SBR. But if it is an RPC operation, A-Shell doesn't support that directly (although you could probably create the equivalent mechanism by running an A-Shell program on the network server that your A-Shell/Linux program talked to.)

You can also execute Perl scripts just by using HOSTEX (under UNIX) or MX_SHELLEX (under Windows) on them, provided Perl is installed. I've given some thought to supporting embedded Perl statements inside of Basic, but I'm not sure it's worth the additional dependencies/baggage (involving dynamic or static links between A-shell and Perl libraries) that that would require.

Check out the socket programming chapter in the A-Shell development guide. I've done socket programming from a Softworks Basic client to a Windows server using Micro Sabio's tcpx subroutine. IIRC I had to send a close connection to get the service (written in Visual Basic) to send it's response. I always felt that was a problem with .net, VB itself, or the VB programmer's code, not TCPX.