RE: Importing XML into FM+SGML

[Dan Emory <danemory@xxxxxxxxxxxx>]

NOTE: This reply is also being posted to the FM+SGML and Free Framers lists.
-----------------------------------------------------------
You want to know whether FM+SGML can import XML document instances.
Here is my analysis:
=====================================================
No version of FM+SGML has the capability to import XML document instances.
In order to overcome this limitation, you must, before importing, convert
XML document instances to valid SGML document instances which conform
to a DTD and an SGML Declaration that must be specified in an
FM+SGML Application Definition. The SGML Declaration must define, for the
FM+SGML parser, how to interpret the markup in the SGML
document instances produced by the XML-to-SGML conversion step.

Of course, you must also develop an EDD (including all required format rules)
from the DTD, as well as a template for import (containing the EDD's element
catalog). The template for import's paragraph, character, table, and 
cross-reference
catalogs must contain all the tags and formatting needed to implement the 
format
rules specified in the EDD.

The FM+SGML Application Definition must also contain all the read/write
rules needed to properly import the SGML document instances produced
by the XML-to-SGML pre-processing step.

Since FM+SGML cannot read Unicode on import, the XML-to-SGML
pre-processing step also must convert all non-ISO Latin1 and
upper-ASCII Unicode characters to the corresponding entity references
in the ISO character sets, making sure that all such entities have a
read/write rule. And, if any of those entities require a code point that
is one of the 5 locked code points in the FM+SGML character set,
you're out of luck. Obviously, Unicode conversion will be impossible
for most non-Latin languages.

If your XML documents use more than one language, then element names
and/or attribute values must identify the language being used in the text 
of each
such element, and the EDD format rules must specify the appropriate character
or paragraph tag (in the template) that will produce the correct characters
in the specified language.

The pre-processing step that converts XML to SGML prior to importing into
FM+SGML will also have to find some way to convert XLinks and
XPointers to SGML-conformant IDREF and ID attribute values.
However, SGML has no way of linking to external docs. Consequently,
if your XML document instances include such links, there is no way
to convert them to SGML-conformant cross-references,
thus, those links will not work after you import them into FM+SGML,.
consequently they won't work in any PDFs produced from FM+SGML.
It might, however, be possible to use the FDK to develop an API that
would make those external links work in FM+SGML.
========================================================
 >We are working on setting up a XML-based workflow. Data will be stored in a
 >database, using XML.. Output must be possible to different publishing
 >systems, an important one is FrameMaker (for a.o. creating printable pdf ).
==========================================================
I believe there are one or more XSL transformation applications available 
which will convert
XML instances directly to PDF, avoiding the necessity of using FM+SGML
for producing PDFs. I know that there was an initiative to develop such
applications, and I seem to recall that one company was offering to
convert XML instances to PDF on a trial basis.
====================================================
 >For this publishing we are creating templates. This far I have no experience
 >with SGML. I am investigating how FrameMaker can Handle XML-data, for
 >creating complex documents. Including tables, Imagecaptions, etc.
=====================================================
The types of complexities described above impose more demands on the
XML-to-SGML conversion and import processes. For example, if, in
the XML document instances, the imagecaption text is contained in a
child element of the graphic element containing the image, then FM+SGML
cannot handle it, because graphic elements cannot have child elements in
FM+SGML. There may also be complications in converting tables that use
the XML table model to the CALS table model used by FM+SGML.  I would
expect that you'd probably need to use the FDK to develop APIs to solve
some of these problems. Either that or you must find a way to do it in the
pre-processing step that converts the XML document instances to
SGML-conformant instances.

SUMMARY:

1. "Well-Formed" XML that does not conform to a DTD cannot be
imported into FM+SGML.

2. DTD-conformant XML must be converted to DTD-conformant SGML instances
prior to importing them into FM+SGML.

3. An FM+SGML Application Definition must be developed, which includes
the DTD, an SGML Declaration, read/write rules, and a template for import
that has imported into it the EDD element catalog. The template must contain
all the paragraph, character, table, and cross-reference formats required to
produce print-ready documents based on the format rules in the EDD.

4. XSL transformations offer the most feasible way to convert XML to SGML
that is compatible with FM+SGML .Such transformations must be capable of
converting Unicode to ISO character set entity references for non ISOLatin1 
and
upper-ASCII characters. However, for central European, Cyrillic, and
other languages, there are locked code points in the FM+SGML character
set which may make it impossible to convert some characters to displayable
characters in FM+SGML.

5. The XML-to-SGML transformation application described in 4 above
must also solve a number of potential incompatiblity issues that
typically arise in complex structured documents. It is likely that some of
those problems will require the use of APIs developed with the
FM+SGML FDK.

6. Unless your XML documents are quite simple, and contain no
external links, the cost of developing the necessary XML-to-SGML
transformation application and the FM+SGML Application Definition
is likely to be high.

7. If the main objective of importing XML into FM+SGML is to produce
PDF output, you should first look into the possibility of an XSL
transformation application that directly converts XML fo PDF right
out of your database.




====================
| Nullius in Verba |
====================
Dan Emory, Dan Emory & Associates
FrameMaker/FrameMaker+SGML Document Design & Database Publishing
Voice/Fax: 949-722-8971 E-Mail: danemory@primenet.com
10044 Adams Ave. #208, Huntington Beach, CA 92646
---Subscribe to the "Free Framers" list by sending a message to
majordomo@omsys.com with "subscribe framers" (no quotes) in the body.



** To unsubscribe, send a message to majordomo@omsys.com **
** with "unsubscribe framers" (no quotes) in the body.   **