boley@uklirb (Harold Boley) (05/17/89)
{This is the slightly updated section 4.4, USER ORGANIZATION, from SEKI
Report SR-88-22.}
The explosive development of XPS shells confronts prospective users with a
bewildering 'zoo' of originals and variants, choices and pressures,
trend and hype.
One useful perspective on shells is that of 'commercialized programming
languages':
(1) Since the main developmental effort was invested by companies --
based on earlier R & D done by universities -- shell concepts are
more often presented in advertising-marketing contexts than in
critical-scientific contexts.
(2) Shell-implementation techniques are not often treated in detail in
publications, since competition encourages confidential and
proprietary procedures.
(3) In their desire to be the best of all, compatibility with other
shell vendors has not (yet) been a widely recognized design goal of
companies.
Although there is a number of journals and newsletters watching the shell
market, these should be complemented by a more interactive way of informa-
tion exchange. Perhaps this moderated UUCP newsgroup on XPS shells will be
a good medium. (It depends on you sending contributions to
shell@uklirb.uucp: A new group provides the best opportunity for switching
from 'being passive information consumer' to 'becoming active in the net'.)
Similarly, no vendor-independent, international "user organization for XPS
shells" seems to exist presently; perhaps it could be formed in connection
with the shells newsgroup. Among other things, it could distribute public-
domain shells (cf. NASA's COSMIC at the University of Georgia), develop
shell benchmarks (perhaps somewhat more complex than NASA's Monkey & Bana-
nas benchmark), organize shell contests (unpublished draft), compile shell
bibliographies (see IRUBESS postings), maintain price/performance comparis-
ons, and work towards a shell standard (see section 4.3).
The field of XPS shells seems to be ripe for some kind of coordination
along these lines. XPSs and shells have become so important in commercial
AI that academic AI must be concerned about their further development,
which may well be critical to the future of AI as a whole.