Self-archiving by authors in the field of mathematics

["Indiana Univ. Math. J." <iumj@indiana.edu>]

Dear All:

I have been mulling over what some of you write for a while now. I could
muse a bit in typical "academese" ("It seems to me that ... might...  
could be viewed as ... but other factors, of course,..."), or I could
speak exactly as I see it, and throw myself to the mercy of you all. I
choose the latter.

Do those who vigorously promote OA have in-depth knowledge of these three
areas:

(1) SMALL, INDEPENDENT publishers' finances,
(2) programs needed to implement some suggestions, such as 
"self-archive your papers" (advice to authors).
(3) "ideological" (philosophical) questions relating to the speed of
publishing versus truly creative work.  (Donald Knuth had something to 
say about this...)

In this email I choose to refer only to (2) above.  My perspective is that
of mathematics publishing, in which I have over 20 years of editorial
experience.

Most mathematicians today use a program called TeX, which comes in a wide
variety of flavors and with a forever-growing number of contributed
packages (macros).  The result of this proliferation of programming
add-ons is that a paper written *today* using LaTeX, say, might not
compile *tomorrow* unless the author went to the extra trouble of listing
the version number of the packages he used today.  (I know this for a
fact: it happened to me.  My experience must have been sufficiently
widespread that the American Mathematical Society offered yet another
programming tool to compile specific package versions into a given paper.)
Naturally, most mathematicians prefer to be mathematicians first - and
leave the vagaries of compiling their paper into a *permanent* fashion to
someone else.  After all, time is limited...

Even if a large fraction of mathematicians could be persuaded to pay
attention to the complexities of LaTeX so as to produce a semi-permanent
version, the question immediately arises that LaTeX itself (like most
other formatters) is in flux: conversions to XML or to PDF are discussed
in terms of which should be the "archival" version.  Better yet, am I the
only one who has experienced that PDF's, by definition a *portable*
format, are not *portable* a 100 percent of the time?  (Rhetorical
question: there are several well-known reasons why PDF's sometimes fail to
be portable.)

Therefore, I am left to wonder whether librarians have lost what used to
be the librarians' greatest emphasis, "Let's safeguard this for future
generations".  A librarian that, today, entrusts self-archiving (of
mathematics) to mathematicians surely thinks that all that is necessary is
to burn a couple of CD's and then, with the passing of time, transfer the
CD's to whatever media should then be in use. Well, that really shows a
total of understanding of technical issues having to do with the fluid
nature of notation in mathematics, and the fluid nature of *viewing*
mathematical writing.  (By analogy, consider a librarian concerned with
the preservation of *books* that said, "All you need to preserve books for
an indefinitely long time is to have them printed on acid-free paper", and
paid no attention to the humidity and temperature of the environment, the
necessity to have copies in more than one location to safeguard against
fires, etcetera.)

Best regards, Elena Fraboschi (Indiana University Mathematics Journal)