electronic publishing

[Robert J. Robbins]

A recent posting has raised some issues with regard to electronic 
publishing and has made particular reference the presentation by Patricia 
Morgan at the last BioMatrix meeting.  The posting recalls two arguments
against electronic publishing, half tones and cost to libraries, and notes 
that less than 5% of computer science journal articles contain photographs.
The posting also notes that, 'when I mentioned the second argument to a 
friend they commented, "who needs libraries anyway?".  That is, we could 
have direct distribution from publishers to readers.'

In response to these comments, I offer some observations:

  -  The posting is correct that the use of halftones is discipline 
     dependent.  Computer science, mathematics, logic, philosophy, 
     and other abstract fields don't need them.  Biology, especially
     molecular biology, does.  Halftones are not the only problem, 
     however.  Other graphics make for difficulty, too.  The argument,
     raised at the meeting, that PostScript can be used as the galactic 
     standard for page layout just doesn't hold water.  First, the 
     standard isn't as standard as standards ought to be.  I have had 
     people send me PostScript reports via e-mail and when I print them 
     out using a PostScript printer at NSF, occasionally a page 
     comes out with a screwed up format.  The source claims 
     the file printed fine at home.  This leads me to believe that 
     not all PS files print identically on all PS printers.  Second,
     not everyone has access to PS printers.  Third, even when folks 
     do have access to PS printers, they often would prefer to receive 
     a printed article in the mail rather than fire up the printer 
     and produce (at, say, 4 cents a page) a 100-page printout 
     that then must be stapled or punched or whatever so that it 
     can be carried around and read conveniently.

  -  The doing-away-with-libraries notion sounds intriguing at 
     first, but appears more naive and silly upon further thought.
     Libraries do not exist just to provide full employment 
     opportunities for parasitic librarians.  They provide real 
     functions, such as allowing researchers access to many more 
     journals than they could possibly afford individually.  They 
     also provide an archival function, which is essential in 
     disciplines that are less ephemeral than computer science.
     Quick, give five computer-science citations that are over 
     fifty years old that can still be read with profit.  However, 
     there are many biological manuscripts that are over a 
     hundred years old that are still important and useful.  Many 
     of these involve detailed illustrations (engravings, usually)
     of anatomical studies that have not been redone, since the 
     original 19th century work is considered definitive.  Converting
     these to PostScript format would be costly and tedious, and 
     would greatly reduce the value of the work.  The archival 
     problem is perhaps one of the more acute arguments against 
     electronic publishing.  Quick, give five examples of material
     written electronically over forty years ago that can still 
     be read easily with current equipment.  For electronic 
     publishing to be considered even as a candidate for archival 
     publishing, there will have to be some possibility that 
     material published electronically will continue to be 
     readable for decades, preferably centuries, without need for 
     periodic copying onto new media in new formats.  

  -  The issue of cost is real, as well, and cannot be dismissed 
     by doing away with libraries.  Publishers publish to make a 
     buck, not to provide a free service to the world.  Serious 
     electronic publishing involves user fees independently of 
     whether the user is a library or an individual.  Serious 
     publishing also involves a concern for the rights, economic
     and otherwise, of the author and publisher.  Sure, you can 
     always make 50 xerox copies of the latest issue of Cell and 
     distribute them to your friends, but that will involve a lot
     of work and expense and you are not likely to do it very 
     often.  Therefore, the publisher of Cell can set subscription 
     fees on the assumption that most readers will be looking at 
     a paid-for copy, not a xeroxed rip off.  With an electronically 
     delivered journal, say that arrives via email, you can, with 
     a simple forward command, send 50 or 100 or 1000 copies off 
     to many friends almost effortlessly.  Therefore, a publisher 
     working in this medium must assume that most of the readers 
     will be using non-paid-for copies and the few fools (libraries?)
     who obtain their copies legitimately must be expected to carry 
     the full cost for the subscription.

I could add a few more observations, but I suspect my position is 
fairly clear.  I believe that the idea that electronic publishing 
will replace print publishing is about as accurate and astute a 
prediction as the one made frequently in the fifties that private
helicopters would replace private automobiles.  Helicopters play 
many important roles in our society, but providing routine 
individual transportation is not one of them.  Likewise, computers and 
electronic communication play important roles in our society, but
replacing the printed word as the primary medium for scientific 
communication is not one of them.

At the same time, there are certain kinds of scientific publishing
that cry out for an electronic medium.  Database materials are 
obviously one of these.  How many people prefer to use GenBank 
in the hard-copy, multi-volume form?  I think that those of us 
who believe that computers have an increasingly important role 
to play in the practice of biology should take care to avoid
making exaggerated claims, either through naivete or excess 
enthusiasm.  Nothing undercuts a good case more than a patently 
false assertion.