RE: Word to Framemaker

["Fred Ridder" <docudoc -at- hotmail -dot- com>]

The first point I would make is that the "WordPad-like text" is
the *real* content of your document. Everything else is really
just presentation, and is either background items (things like
logos, titles, and dates that appear in the headers and footers
for page after page), content that is derived programmatically
from the *real* content (TOC, running headers/footers that
are derived from headings, index, the text portion of cross-
references), and purely presentational content that is
generated by the tool itself (page numbers, autonumbering,
autotext, bullets).  Word basically munges all of these into
a single layer (particularly in Print Layout view), while
FrameMaker separates the real content from the formatting
layer (the paragraph and character format catalogs), from
the background layer (master pages), and from the derived
and system-generated content (system variables, generated
files.  (A structured authoring environment using XML or SGML
goes even farther in separating content from formatting and
presentation, but that's another whole discussion...)

In our group's migration from Word to FrameMaker, I was
responsible for converting dozens and dozens of documents,
totalling >10K pages, and in my experience the key to success
was leaving behind as much of the Word-specific presentational
features as possible.  Headers and footers were already
pre-built in my FrameMaker templates. Autonumbering and
bullets and autotext for notes and cautions were pre-built
in my FrameMaker templates, so I wrote Word macros to strip
all that formatting out of the Word docs. In FrameMaker we
were going to insert graphics by reference, so there was no
point in trying to convert the in-line graphics in the Word
docs; a macro got rid of them pre-conversion. In FrameMaker
we were going to use a file-per-chapter model rather than
our mnolithic Word docs, so cross-references were inevitably
going to get broken; another Word macro converted all x-refs
to plain text and marked them with a named character format
so that we could find them and reconstruct them post-
conversion. Index entry fields in the text came across just
fine (I don't know why the are not doing so for you); all I
had to do after conversion was add the generated file to
the FrameMaker book and click the Update Book button.
ToC, LoF, and LoT were similarly trivial to rebuild after
conversion using pre-formatted templates; it took less
than 10 minutes per book to completely reconstruct all
the generated files. The slowest part of the conversion
and clean-up was dealing with tables, since Word and
FrameMaker have fundamentally different underlying models.
But Rick Quatro's wonderful TableCleaner plug-in takes a lot
of the pain out of with some really intelligent global
operations.

My opinions only; I don;t speak for Intel.
Fred Ridder
Intel
Parsippany, NJ