You might be interested in the opinions expressed
by the a MIT Press guy on the subject of
of electronic publishing vs. paper publishing.

					-BB


> 
>   In a recent review of a textbook about high-performance computing for
> undergraduates in science and engineering ("An Introduction to
> High-Performance Scientific Computing," by Lloyd D. Fosdick of the University
> of Colorado and three colleagues), I commented on the difficulties that
> publishers encounter in trying to keep up with the rapid developments in
> fields like HPC. I noted that inertia in the publishing process adds to these
> problems. (See HPCwire article 8719 REVIEW: NEW MIT TEXT VALUABLE BUT ALREADY
> OUTDATED, 06.14.96)
> 
>   My remarks were, I must admit, biased by my own experience as an author
> and by complaints by other authors. It is well known that the word "author"
> can be defined as "a person highly protective of his or her own creative work
> who is instinctively hostile toward a hereditary enemy, known as a
> publisher."
> 
>   I am indebted to Michael Rutter, an acquisitions editor at MIT Press, which
> published the book under review as part of its very useful series on
> scientific and engineering computation.
> 
>   Rutter offers the publisher's point of view. His remarks are excerpted
> below.
> 
>   This a matter of common-sense fairness. His explanations should also be
> useful to HPCwire readers who are contemplating writing a book as well as
> authors already enmeshed in this process, professors seeking the most
> effective materials for their courses, and the students who will shape the
> future of HPC into the 21st century.
> 
>   Rutter said:
> 
>   "I just wanted to quickly respond to your comments about textbook
> publishing.
> 
>   [You noted that] "'Many textbook houses customize collections of
> traditional textbook materials, journal articles, excerpts from anthologies,
> and other material to suit professors' teaching plans. In this process, which
> might be called Kinko publishing, the customized text is assembled, printed,
> and delivered within a few weeks or less.'
> 
>   "That is definitely the case. DocuTex from Xerox has significantly
> improved the quality of the 'make-your-own' textbook process. But it is
> prohibitively expensive (often costing about 2 or 3 times more than
> traditional printing -- i.e. the film method).
> 
>   "Further, there is the issue with permissions [to reproduce copyright
> material] (also very expensive). And finally, being a recent student
> subjected to such textbooks, I have to confess that they are very difficult
> to use. A consistent, single voice (no matter how narrow) time and time again
> proves a better teacher, than a blizzard of materials by a myriad of authors
> (from a host of different time periods) loosely linked together.
> 
>   "A good text often fills the gaps where a professor is least proficient ...
> and allows a student to have a tutor away from the classroom."
> 
> DO COMPUTERS SPEED UP THE PROCESS?
> 
>   Rutter also noted that I had commented: "Traditionally, the delay between
> the submission of a manuscript and its publication was justified because of
> a time-consuming transformation that required type-setting, proof-reading,
> and reproduction of any charts, graphs and illustrations. Computers have
> shortened this process. This book, for example, was set in Latex by the
> authors. Thus the publisher could skip the time (and expense) of a large part
> of the production process."
> 
>   Rutter replied: "...Certainly, type-setting is changing, especially with
> LaTex (a good math equation/computer science code editor, and terrible
> type-setter), and other formatting programs.
> 
>   "But, here's the catch. Most modern vendors have more difficulty outputting
> LaTex files then they do creating a book from scratch. Surprisingly, it often
> takes more time (& expense) to successfully output LaTex files then it does
> to re-typeset the entire book.
> 
>   "I know this is tough to believe -- but we (and other publishers) have been
> through this process since LaTex was invented. Its promises have not panned
> out."
> 
> NOT PLATFORM INDEPENDENT
> 
>   "Postscript -- as much as it is claimed to be -- is not platform
> independent. Inevitably, something 'strange' happens on author A's machine
> when he/she converts LaTex files to .ps files [causing] problems later on.
> Almost none of our vendors want to work with source or dvi files.
> 
>   "Why? Often an author hacks his/her own macros ... and that can be a mess
> (even when we supply them with our own -- they often usurp us by creating new
> routines). Some of the technical problems are the fault of the vendors --
> they are still playing catch-up.
> 
>   "Most authors who think they are typesetters are not. Typesetting is a
> tough job -- merely having a Mac and a mouse doesn't command authority or
> expertise.
> 
>   Most authors who think they have proofed their manuscript have not done a
> proficient job. Proofing consists of a lot more than simply checking the
> spelling, and punctuation. It takes years of practice to be good at it."
> 
>   Rutter also quoted me as saying: "The rest of the delay can be
> attributed to the institutional inertia that still afflicts too many
> publishers -- even in technical fields."
> 
> IS THE WEB AN APPROPRIATE MODEL?
> 
>   He replied: "I think that is a broad and false generalization. We (and
> other presses) are constantly trying to adapt to the new on-line world (with
> course support materials, software, and even entire texts, all available on
> the Web).
> 
>   "Soon you will see MIT Press's latest electronic effort -- a large project
> that should be up in about six months. Watch for it.
> 
>   "Academics are still frightened about publishing material on the Web. They
> feel that it doesn't have the same cachet as traditional books -- and for the
> most part, they are right. Much of the reluctance comes from them, not the
> publisher.
> 
>   "I agree that textbook publishing will change in the future (I am not a
> luddite by any stretch of the imagination). The Web opens up a great number
> of possibilities for both students and publishers.
> 
>   "Traditional, textbooks -- good ones that are used year after year -- take
> years to develop. Creating a textbook is not merely just slapping some
> random bits together and updating it on the fly. For certain fields, Web
> publishing might prove perfect ... but not for all fields. Textbooks
> consolidate, as best they can, material in a field. They are never completely
> up to date, but no one (student, professor, etc.) can possibly keep track of
> everything that is going on in a given field.
> 
>   "You need a ground to stand on somewhere ... to create an amorphous
> book would be a horrific way to learn anything. Linear learning is not a bad
> thing -- it has worked for centuries.
> 
> RANDOM CHAOS@WWW.LOW QUALITY.html
> 
>   "If you want textbook publishing to look as random, and chaotic as the Web,
> and to equal it in content quality (incredibly poor on a consistent
> basis), then take away the publisher and the 'book'. It seems that with more
> and more information out there, there is more and more need for
> publishers/editors to guide readers through it.
> 
>   "I do not think this is a mere argument for self-justification, [caused by]
> fear that the publisher will go away. I do not think that the publisher will
> disappear, but rather, will change.
> 
>   "I think a lot of people are still too optimistic about the Web and believe
> that merely putting something 'up' makes it 100% better -- even though it is
> tough to read, hard to access, and not organized that well. People are buying
> more and more books these days (and strangely, buying them through Web
> sites).
> 
>   "Again, I enjoyed your review ... and I am glad you wrote it. But I
> think we are far away from creating the perfect textbook."
>