These propositions speak only to part of what the Committee seems likely to
recommend -- we'll need to develop similar sets of propositions and
recommendations for the expanded-commmunity, multimedia development, and other
possible areas we've discussed.
gj
e40-359a/MIT/Cambridge MA 02139
voice: (617) 253-3712
fax: (617) 258-8736
url: http://web.mit.edu/user/g/j/gjackson/www/home.html
key: pgp@pgp.mit.edu
0. To be realistic, we must assume that there never will be substantially more
funds for MIT academic computing than there are now, and that continued slight
decline is likely, so that doing big new things requires current things to
change.
1. The fastest-growing demands for educational technology are in two areas:
(a) communication and information based on Web services and other Internet
tools, and (b) applications that require greater computational and graphical
power than we can now afford in basic Athena workstations. Requests for more
powerful machines have now come at least from Chemistry, Chemical Engineering,
Materials Science and Engineering, Architecture, Urban Studies and Planning,
Aeronautics and Astronautics, Biology, Literature, Economics, and a few
faculty in other departments. These requests are part of the impetus behind
the SGI project, but in general they can't be met given the minimal space
between our resources and the demands for academic-computing services. Even
though demands for more powerful computers are growing, demands for
traditional computers aren't shrinking. We can't simply divert resources from
"traditional" to "powerful" computers without making other changes.
2. Basic computing and network services have become central enough to all
students' education at MIT that each student should own a computer. This
should become a requirement forthwith, with a limited set of machines
specified to meet the requirement (probably a given level of Windows machine,
a given level of Mac, and some appropriate specification within the Unix and
perhaps the DOS categories). Were all undergraduates to purchase new $2,000
computers at entry and amortize them over four years, the total cost
aggregated across all undergraduates would be about $2.6m; if MIT provides
grants to cover about 1/3 of undergraduates' educational costs (which I think
is about right -- 55% of all undergraduates receive financial aid, but not all
receive grants from MIT general funds), this implies that a computer-purchase
requirement would increase MIT financial-aid costs by about $900k annually.
3. If all students own computers and have network drops, then most of what's
necessary to use Web and other Internet tools is in place. The only additional
requirement is that faculty and others involved with MIT education be
networked as well, and to this end financial and other barriers to network
connectivity for faculty should be removed by having central budgets, such as
the Deans' or the Provost's, pay the costs of that connectivity. Without
counting discounts and non-participants, the annual $300 costs of network
drops for 1,000 faculty total $300k; if half of all faculty currently lack
MITnet connections, the one-time cost of connecting them (again ignoring
volume discounts and such) is $150k. In the inimitable words of one committee
member (not me!), finding this much money in a $300m instructional budget to
network faculty "...should be a no-brainer".
3. If all students own networked computers, it becomes possible to outlaw or
reduce in priority certain widespread, time-consuming current uses of public
workstations. For example, reading newsgroups, doing recreational Web surfing,
and playing games might become unacceptable in public clusters, and working on
documents, reading mail, doing basic calculations, or simple programming might
become low-priority uses for public machines (that is, students doing these
things would have to give way to students with higher-priority needs). Of
course there would continue to be some public machines available for basic
functions, for example so that fraternity dwellers could read their mail
during the day, but these might have very short use limits analogous to those
on the Express machines. (We might paint Express machines red.) There might be
"docking" stations for mobile computers.
4. A reduction in basic document processing, programming, and network-service
use of public Athena workstations should permit us to reduce dramatically the
number of "Athena" seats devoted to traditional, generic workstations capable
of everything, perhaps by about half.
5. We can then replace the no-longer-necessary half of the traditional Athena
seats with fewer but specialized machines, which may mean more expensive and
powerful machines (graphics, intensive computing, multimedia, and so on). The
total cost of specialized machines would have to total no more than the total
cost of the no-longer-necessary machines. Thus, for example, if specialized
machines cost on average twice what a traditional Athena workstation currently
costs, as is the case for the Indys we're deploying experimentally, then the
total traditional+specialized seats in the "Athena" environment would shrink
by 25%. Of course it may also be possible to reduce the cost of the remaining
"traditional" machines, and the resulting savings could be invested in
specialized equipment.
6. Making the Athena environment heterogeneous in this way will introduce
numerous complexities including support, development (not necessarily full
ports, but certainly basic services for specialized machines), scheduling and
access, operations and maintenance, software licensing, renewal and
replacement, departmental partnerships, and so on. This may consume additional
resources, and further reduce seats.
7. Because of financial-aid implications and MIT's general unwillingness to
impose requirements retroactively, all this must happen in stages. It's
probably possible to begin requiring entering classes to own computers quite
soon (by fall '96, say), and to do networking things for faculty on a similar
timeline. The rest of these changes will then unfold over about the time it
takes the affected entering classes to move through the Institute (say two or
three years).