[1518] in Public-Access_Computer_Systems_Forum
Citations to e-files
daemon@ATHENA.MIT.EDU (PEMMONS%WCU.BITNET@mitvma.mit.edu)
Mon Oct 19 16:48:30 1992
Date: Mon, 19 Oct 1992 15:25:19 CDT
Reply-To: Public-Access Computer Systems Forum <PACS-L%UHUPVM1.BITNET@mitvma.mit.edu>
From: PEMMONS%WCU.BITNET@mitvma.mit.edu
To: Multiple recipients of list PACS-L <PACS-L%UHUPVM1.BITNET@mitvma.mit.edu>
In-Reply-To: Your message of 10/16/92 10:54
----------------------------Original message----------------------------
You have touched, I think, on an issue of profound significance not only
for scholarship but for society. As to officially published sources
(electronic counterparts to frequently updated reference books which might
be cataloged as quasi-serials), solutions may not yet be at hand but are
usually feasible. Any contribution your research can make towards some
standard citation procedures and formats will be valuable; I am sorry
that I don't know my way around well enough to contribute to your project with
any helpful specifics.
A deeper question, however, is to what extent it is in the nature of
electronic data to assume either a permanent form or a form in which changes
are emitted at discrete intervals of time. Historians have already expressed
some alarm over the fact that the move towards paperless offices deprives
them of source material for their work. Needless to say, the opportunities
for malfeasance run rampant if there are no audit trails or if an electronic
resource represented as permanent undergoes surreptitious modification.
We all heaved a sigh of relief when 1984 came and went and we still considered
ourselves possessed of coherent past, present, and future. But wait awhile...
The news isn't all bad, insomuch as a person who has captured two versions
of a file can employ computer utilities to find the differences. But this
presupposes having two versions, and finding differences between them may
not equate to determining when changes occurred or who made them.
Software development faces this problem in control of constantly evolving
source code. No traditional bibliographic methods, such as labeling
successive files as to "edition" (a.k.a. version numbers) or "Publication
date" (a.k.a. date/time stamp) suffices for the problem in any but the
crudest sense. Version control demands deriving the state of
the file at any desired time by use of (1) A snapshot of its state at
one time (2) A record of the changes which have been/will be made to it.
This technique has the advantages of accuracy, flexibility, and usually
of compactness; especially inasmuch as one is often more interested in
the modifications themselves and when they were made than in the data as
a whole.
What we will need, in other words, before the bibliographic citation of
electronic data has become a technique with any realistic conformity to
the idiom of its material, is to handle delta, change-to. Bibliography
must move from algebra to calculus. Changes made to a data entity must
be honestly kept and disseminated along with a state of the entity at a
single time.
It is a problem which, I fear, will be made insoluble in exactly the
cases where a solution would be most desired.
