The former HTML version of this document was referenced in Issue 22 of Linux Gazette in the article Word Processing vs. Text Processing?
The author of that article and I see fairly eye-to-eye on the fact that there are differences between the generic term "Word Processing" and the function that people actually need to perform in "creating a document."
An article that provides another perspective (e.g. from that of a professional typesetter) on this is the Seybold article, What has WYSIWYG done to us?
I received email from one person with a dissenting opinion that felt that my "mission" to separate out the various aspects of document creation as I do represents a creation of unnecessary complexity. My response to that is thus:
There will surely be cases where the overall process of splitting out the various document abstractions is overkill. If all you want to do is to write a simple letter, to format it with some form of physical markup, and print and send it, then there is little value to analyzing the different "layers" of abstraction, meaning, and processing.
You hardly need to compose a DTD to write a simple letter.
By the same token, you hardly need a computer to write a simple letter. A typewriter will suffice, and the modern "electronic typewriters" are highly likely to be highly suitable to this task. There is relatively little benefit in using a computer for this, and if you do, you might as well use as simple a program as possible.
I'm not very interested in such simple documents. There are quite enough documents that need to be written that are sufficiently complex as to require things like SGML to manage their structure.
The average book on a bookstore shelf is often complex enough; technical books are almost always so complex as to require a good deal of sophistication. The Linux Documentation Project is an extremely good example of a project that needs the most sophisticated documentation tools (such as SGMLTools ) that we can get our hands on. When documenting complex software systems, complexity is a given, and we need to bring to bear on this the best tools that we can.
Other analogous "rants" may be found in various places:
A document on why writers might want to consider using Emacs to edit their documents.
This SIAM News essay discusses ways in which the handling of mathematical material has (and, to a considerable degree, has not) progressed over the last thirty or so years. It observes that TeX represented a significant advance, and that, in many ways, there has been little advance beyond. Computers have gotten faster, which, for a time, led to advances in document processing times through the 1990s, but changes have somewhat stalled since.