I'm in need of a library, application or generally-applicable source
code to allow me to dynamically produce a valid Microsoft Word 97 .doc
file on a Linux system.

I need to be able to convert a document with arbitrary markup (a
formatted text file, an HTML file, whatever) to a Word document
containing font definitions, formatting (bold, italic, strikethrough,
etc.) and custom styles according to a custom Word template.  I can
preprocess the original file to a special format for the converter if
necessary.

Using Office automation components is not an option.  I've already
looked at wv, antiword, libole2 and other similar applications and
libraries, but they all appear to only support importing.

I know this is possible, as applications such as AbiWord and
OpenOffice.org Writer convert from their internal XML-based formats to
Word, but taking apart these applications is not an option, and their
exporters do not appear to be libraries.

An answer providing links to potentially usable software, including
link to the exporting specifications of that software, would be most
appreciated.  Libraries that are Free software are preferred,
generally open source is fine, and existing Perl, Python or PHP
bindings are a plus.  Commercial libraries will be considered only if
nothing else is available (and I don't believe that is the case).

---

Request for Question Clarification by rhansenne-ga on 06 Jul 2002 10:03 PDT

Hi,

Is conversion to .rtf/RichText files (which can be easily read and
converted by Word) an option or does the output file absolutely need
to be .doc format?

Kind regards.

---

Clarification of Question by vito-ga on 06 Jul 2002 10:33 PDT

The output absolutely, positively, has to be a valid Word 97 .doc
file, supporting fonts, formatting and custom Word template styles. 
RTF is not an option, or I wouldn't need Google Answers.  :)

Hello vito,

In this link (http://www.w3.org/Tools/Word_proc_filters.html) you can
find a compilation of word processor filters.

Of the filters described in the link above, it would appear that
Filtrix (http://www.blueberry.com/index.html) from Blueberry Software
would meet your requirements.  It costs $189 for a Linux license.  See
http://www.blueberry.com/ldomof.htm.  Also, it costs $495 for a fixed
Unix license and $795 for a floating Unix license.  See
http://www.blueberry.com/udomof.htm.  The import/export capabilities
of the Unix version of the program can be found in this link
(http://www.blueberry.com/pisu.htm).  A free demo of the program is
available.  The demo for the Linux version of the program can be
downloaded from here http://www.blueberry.com/ftpall.htm.

In the alternative, you could try to write your own conversion
software.  A document describing Word 97’s Binary File Format can be
found here http://www.btinternet.com/~shaheedhaque/generator_wword8.htm.

Using the AbiCommand plugin with AbiWord, you can supposedly automate
Abiword from the command-line.  You may very well be able to use this
to perform the dynamic conversion you need.

It's available on abiword's plugin page in rpm format.

Abisource:  http://www.abisource.com/
Plugin page:  http://www.abiword.com/download/plugins.phtml
Actual File:  http://unc.dl.sourceforge.net/sourceforge/abiword/abiword-plugin-abicommand-1.0.2-1.i386.rpm
Another (simpler) approach that may work is simply using the "-to"
command-line argument to have abiword automatically convert the file. 
However when I tried this executing, "abiword -to doc testfile.txt"
the resulting file was actually formatted in rtf.  Though I am running
Abiword version 0.9.5 and the newest version is 1.0.2, so you may have
better luck.

Vito -

You may want to check out the Jakarta POI project
[http://jakarta.apache.org/poi/]  it is centered around
reading/writing documents created in the OLE2 document format using
Java.  The Excel piece of the project is fairly mature, while the
Word/Doc part of the project is in the early stages (though I have had
success using the code to *read* Word documents on Unix with Java). 
They also have links to other, similar efforts.