On Tue, 27 Apr 2004, Peter Bex wrote:

> I've tried several things, but they all aren't real solutions:

> - Create PDF with LaTeX and ask Windows users to install PDF Creator.
>    Works perfectly for one-way communication, but when changes are required
>    from the other party, this simply doesn't work.

Actually, I think it's a bit dangerous to let your group members directly
edit files.  I once was working on a project (using MS Word) in a computer
lab, with all of my group members present.  I stepped out for a few
moments and when I returned, a good deal of my formatting was ruined,
because someone had decided they could "just make a few changes" while I
was away.

Even when receiving changes in .doc format, I've often found that the
changes are often unusuable as-is; people tend to change more than just
the text, and I have to go back and fix fonts, sizes, etc.

In short, appointing one person to be the "document manager" works best.
And, the tools that he chooses to do his job shouldn't much impact the
other group members.

> - Use plaintext and add markup just before printing the final version.
>    Not very comfortable, tables are a bitch to convert and this way I'm just
>    confirming the idea that Unix (or NetBSD) is old-fashioned and only for
>    stubborn people.  And that's always a Bad Thing(TM) to have people think.

For the past few (group) papers I've done, this is the way I've accepted
submissions.  People can use whatever they want to type the document, and
then save it to RTF or plain text (without any images).  They e-mail me
the file, as well as any images they'd like to insert (and where to insert
them).  This avoids that "RTF bloat" that you mention.

I then import their work into a LaTeX document.  If they want to make
further changes, they can e-mail me a list of changes to make (or, they
can e-mail me paragraphs of text to re-import).

I used to dread getting large tables, but I've been using the 'tabularx'
package with my own columntypes and found it to work quite well.  (Though
I think tables with big chunks of text really defeats the purpose.)

Usually, there are some initial complaints that I should be using Word
(etc.) but the complaints seem to stop when they see how I renumber all of
the references, figures and tables, and generate the ToC.  (I type "make"
and a new PDF spits out.)  Watching their jaw drop is a rewarding
experience, as is seeing other groups spending too much time on this step.

-- 
Tyler Mitchell