RE: Future of FrameMaker: InDesign?

[David Cramer <dacramer@xxxxxxxx>]

>David's question as why the Frame MIF format is not widely applicable
>focuses directly on the reasons why XML is having such a fundamental impact
>on documentation technology.

Yes, I agree. I guess one of my fears is that XML markup will be 
unusable for many format oriented products. I have often noticed a 
dismissive attitude of SGML/XML advocates toward formatting in 
general. If all printed/webbed output were to be represented 
accurately in a Venn diagram, I believe the circle for 
format-oriented material would overlap the one for 
content-object-oriented material; but some of the SGML/XML proponents 
seem to want me to believe that the content-object-oriented circle 
would contain everything. Oh well, wishful thinking motivated by a 
desire to market one's SGML/XML skills or product are understandable 
;-)


>Personally, I have never had any reason to explore the MIF possibility.
>However, my off the cuff understanding (and others should correct me if I am
>wrong) is that MIF is an proprietary but openly (i.e., published) standard
>markup for describing FM documents, in the same way that RTF is an open
>standard for MS Word documents.

After comparing both standards, I would agree only if "in the same 
way" is taken awfully loosely. MIF seems to be intended to be read 
and used by humans as well as by a program. RTF is not so intended, 
and hasn't been all that standardized from what I have seen. MIF is 
enormously more stable. And its development and publishing were not 
corrupted by the motivations of a company bent on world domination. 
(OK, OK, back on topic!)


>Such standards are still format oriented and relate to and work best with
>the logic of the proprietary formatting engine. Thus they represent the
>document as it is formatted in a paper paradigm. Also, given that format
>oriented systems offer a number of different ways to apply similar formats,
>it is likely that MIF is considerably more difficult to interpret than pure
>SGML/XML. This is certainly the case for RTF.

Dan and others are experts in MIF, so I would appeal to their 
opinions on this; but my impression, having successfully created a 
HyperCard stack around 1994 to dismantle a MIF file into its 
component parts, is that MIF is pretty easy to interpret. So its 
limitation is that it is format oriented. Actually, it might be more 
accurate to say it is object oriented; but the only objects it can 
know about are format objects ;-)

>
>By contrast SGML and XML are international standards for marking up the
>logical structure of documents. The document type definition (DTD) defines
>the different elements of content allowed to occur in a document conforming
>to that DTD and sets rules defining where particular elements may (or must)
>occur within the overall sequential and hierarchical structure of the
>document. Except in special cases like the HTML DTD, which was defined
>specifically to express formats, SGML/XML documents convey only structural
>information, and leave it up to independent processes to apply formats. The
>standards are maintained and published by international bodies including
>representatives of a wide range of application developers and major end-user
>organisations to ensure that the markup languages meet specified goals.
>
>For this reason SGML/XML markup is ideally suited for use by information
>management and processing applications which need to understand the logical
>and semantic structure of documents. Because the primary concern of MIF and
>RTF is format, there is far too little control over how the format is used
>for the markup, so these markups are of little use in logical or semantic
>processing of document content. Based on my own experience, with properly
>designed styles and very well disciplined users, you can readily convert RTF
>to semantic markup (SGML/XML), but you have to have fairly powerful demands
>from users not to do your work in SGML/XML in the first place.
>
>In this sense, MIF is no better than RTF. Basically both markups provide
>published interchange standards which allows developers to build
>applications able to port in and out of the respective internal formats, but
>neither comes close to the planned universality and structural controls
>provided by valid SGML and XML.
>
>In a sense, this is another version of the application holy wars arguments.
>Word, FrameMaker, RTF and MIF represent documents as formatted text on paper
>- good only for reading by humans. The SGML/XML applications represent
>documents as logically structured elements content which can be readily
>parsed and processed by computer systems, as well as being able to be
>formatted for the human users.

It's the process of ensuring that the structured content actually 
gets formatted for human users that I fear may be being glossed over 
here. In practical terms, a lot of documents (10%? 20%? 60%) are good 
only for reading by humans, simply because the process of determining 
and canonizing their structure is more trouble than it's worth. And 
yet most humans have no trouble making use of those documents in 
their unstructured state.

But that's a quibble. The structural approach is probably appropriate 
for the kinds of documents we are considering.

So as far as I am concerned, then, the real point should be how 
quickly, if at all, can Adobe evolve FrameMaker into a wonderful 
formatting front-end to structured documents, seamlessly integrated 
into a total end-to-end process from content objects to formatted 
objects (and back?), and always giving the end-user (whichever end 
they're at) freedom as to how much enforcement want over the 
structure they control.

What puzzles me is that it would seem that a majority of formatting 
challenges are already pretty well handled in a published, stable, 
platform-independent, robust formatting standard like MIF by a 
relatively stable, robust, platform-independent application like 
FrameMaker; so why am I hearing so little interest in using MIF or 
FrameMaker in the future? Is it simply frustration with Adobe for not 
responding to repeated complaints about FrameMaker's shortcomings? Or 
is there something inherently flawed with MIF as a formatting 
description standard?

It seems like a great idea to be able to search more effectively for 
objects using semantic markup. But that assumes there is a 
meaningful, affordable method for getting enough of that content into 
a structured form and marking it consistently enough to be useful. 
Call me a skeptic; but I think there's a little too much XML 
*marketing* and not enough *markuping* in the proposed scenario.

Respectfully,

David
-- 
David Cramer, Process Innovation Evangelist          87-1313 Border Street
PBSC Computer Training Centres (an IBM company)      Winnipeg MB R3H 0X4
Corporate Office Research & Development              Canada


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