Re: EDD Format Design Approaches

[DW Emory <danemory@xxxxxxxxxxxxxxxxxx>]

1. Eric Dunn extols the advantages of format change lists. So do I.  In my 
earlier post entitled "Documenting EDD Design" I stated:
==================
Clearly, an EDD implementation in which all (or at least most) 
context-based EDD format rules specify format change lists (all of which 
are typically located at the end of the EDD) is the optimal approach. This 
allows a single format change list to be referenced from many different 
format rules, and also permits using format change lists as building blocks 
by combining several of them to define a particular format in a particular 
context. By placing all of the format change lists at the end of the EDD, 
you can easily create multiple versions of a single EDD simply by modifying 
the format change lists, leaving everything else unchanged.
============================
2. Eric argues that Approach 2, where only one (or a very few) Element 
Paragraph Format tags are specified, and elements inherit those tags (plus 
any format changes) from their antecedents is not a bad thing, even though 
it denies the template designer virtually any ability to modify formatting. 
His fundamental argument is that the EDD designer and the template designer 
must be the same person, and all formatting decisions should be made by the 
EDD designer in the EDD format rules. Thus, once the EDD is completed, a 
separate and independent template design activity cannot exist. My 
arguments against Approach 2, and in favor of Approach 3 are many and 
varied, and include:

A. Formatting requirements inevitably change over time for all sorts of 
reasons. As long as there is no change in structure, EDDs should be 
invulnerable to change. An Approach 2 EDD is extremely vulnerable to 
changing format requirements. An Approach 3 EDD is not, because it cedes to 
the template designer those formatting parameters which are most likely to 
change over time, as well those which must be changed for different 
deliverable types and different customers.

B. Supose, for example, there is a requirement that review copies be 
double-spaced? Under Approach 3, the template designer would simply create 
an unstructured review copy template containing all the EDD-specified 
paragraph tags, and change the line spacing to double-space. This can be 
done under Approach 3 because line-spacing is ceded to the template 
designer . The revised paragraph formats are imported from the review copy 
template into a copy of the source document to produce the review copy. 
Under Approach 2, you'd have to create a separate version of the EDD, 
making extensive changes to the format rules/format change lists in order 
to accomplish this simple step.

C. Perhaps the most powerful argument in favor of structured documents is 
that it greatly enhances the long-term preservation of legacy documents, as 
well as the capability to reuse information chunks from such documents many 
years after they were created. This tenet assumes that structure has a long 
lifetime. But formatting is short-lived. Thus, if the EDD format rules 
control all formatting, the life of the EDD (and the structure it defines) 
is foreshortened. In other words, the life of the EDD lasts no longer than 
the EDD designer's original vision of how information should be formatted. 
Under Approach 3, any structured legacy document (or a reusable component 
of that document) can be converted to the current formatting standards 
simply by importing the new paragraph formats from an unstructured template.

D. Once an EDD has been created, vetted and formally adopted, there is 
(usually) great resistance to changing it, because doing so would (among 
many other things) erode the preservation of legacy documents already 
created from that EDD. Thus, if the EDD was created under Approach 2, the 
need for changing formatting will be overridden by the resistance to 
changing the EDD. An EDD developed under Approach 3 eliminates the 
necessity to alter the EDD when formatting changes are required.

E. In many cases, EDDs are designed by outside consultants or itinerant EDD 
developers (a recent post on this list desperately seeks an EDD developer 
in New York for a 3-month assignment). Even if the EDD was developed by a 
company employee, you're relying upon the EDD originator to still be around 
when format changes to an Approach 2-type EDD are required. If the original 
EDD developer is no longer available (or offers to make the changes at a 
price you can't afford), you're up the proverbial creek. A well-documented 
EDD developed under Approach 3 shifts the burden of required format changes 
to a much more available (and less costly) talent--template designers.

F. Under Approach 2, what happens when an edict comes down requiring major 
changes in the formatting of corporate or customer documents? You have to 
change the EDD, that's what happens. Under Approach 3, you simply have a 
template designer modify the affected paragraph formats in the template.

G. Inserting voluminous comments in an EDD is not a sufficient solution to 
documenting an EDD design. That approach is analogous to documenting 
complex computer programs solely by comments within the code. Technical 
manuals are needed so that users can understand how the program works, and 
how it can be utilized. In most cases, EDDs are poorly documented, and 
rarely provide a sufficient overview of the design concepts. A complex EDD 
developed under Approach 2 is like computer spaghetti code--it is extremely 
difficult for anyone other than the original EDD designer to modify or 
understand it. Modifying the formatting for a particular element may 
produce unintended consequences in the inheritance of formatting by that 
element's children, grandchildren, great-grandchildren and so on. Testing 
such EDDs to check for, isolate, and correct such formatting anomalies can 
be extremely difficult, even when the original EDD designer is involved. An 
EDD developed under Approach 3 is much easier to understand and modify 
because it eliminates format inheritance. Since an Approach 3 EDD cedes to 
the template designer those formatting parameters which are most likely to 
change, it is relatively easy to document such an EDD in a manner that 
clearly describes how the template designer can successfully modify 
document formatting without affecting (or being affected by) the EDD.

FrameMaker/FrameMaker+SGML Document Design & Database Publishing
DW Emory <danemory@globalcrossing.net>


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