/FONT#0=BASL30/FONT#1=BASI30/FONT#2=BASB30/FONT#14=NGR40/FONT#3=NGR25/FONT#4=NGR20/FONT#15=STA200
␈↓␈↓↓␈↓α␈↓∞␈↓β␈↓∧␈↓∂␈β↓↓␈↓∂S␈↓∞
␈↓ ↓H
␈β↓Q␈↓ ¬	STANFORD UNIVERSITY
␈↓ ↓H
␈↓β␈↓ ¬_STANFORD, CALIFORNIA 94305
␈↓ ↓H
␈↓∧
␈↓ ↓H

␈↓ ↓H

␈↓ ↓H

␈↓ ↓H
ARTIFICIAL INTELLIGENCE LABORATORY␈↓ 
+   Telephone:
␈↓ ↓H
COMPUTER SCIENCE DEPARTMENT␈↓ 
+415-497-4430
␈↓ ↓H
␈↓␈↓ ¬k14 January 1975
␈↓ ↓H

␈↓ ↓H

␈↓ ↓H

␈↓ ↓H

␈↓ ↓H
Professor Alphonse Juilland
␈↓ ↓H
Department of French and Italian
␈↓ ↓H
Stanford University
␈↓ ↓H
Stanford, California 94305
␈↓ ↓H

␈↓ ↓H
Dear Alphonse:
␈↓ ↓H

␈↓ ↓H
␈↓ α_Here␈α∞is␈α∞the␈α∞letter␈α∞you␈α∞asked␈α∞for␈α∞on␈α
administrative␈α
expenses.␈α
If␈α
I␈α
had␈α
hope␈α
that␈α
it␈α
would␈α
be
␈↓ ↓H
␈↓ ↓Heffective,␈α
and␈α
more␈α
time,␈α
it␈α
would␈α
be␈α
better␈α
written:
␈↓ ↓H

␈↓ ↓H
␈↓ β{ADMINISTRATIVE EXPENSES AT STANFORD
␈↓ ↓H

␈↓ ↓H
␈↓ α_Many␈α⊃faculty␈α⊃have␈α⊃come␈α⊃to␈α⊃believe␈α⊃that␈α⊃the␈α⊃ratio␈α⊂of␈α⊂$15,000,000␈α⊂for␈α⊂academic␈α⊂salaries␈α⊂to
␈↓ ↓H
␈↓ ↓H$21,000,000␈α∪to␈α∪administrative␈α∪is␈α∪not␈α∪optimal␈α∪for␈α∪Stanford's␈α∪educational␈α∪program,␈α∩and␈α∩that␈α∩the
␈↓ ↓H
␈↓ ↓Hprojected␈α∂increases␈α∂of␈α∂these␈α∞numbers␈α∞to␈α∞21␈α∞million␈α∞and␈α∞31␈α∞million␈α∞respectively␈α∞over␈α∞the␈α∞next␈α∞five
␈↓ ↓H
␈↓ ↓Hyears␈α
does␈α
nothing␈α
to␈α
rectify␈α
the␈α
imbalance.
␈↓ ↓H

␈↓ ↓H
␈↓ α_I␈α
agree␈α
and␈α
I␈α
have␈α
some␈α
specific␈α
points␈α
and␈α
some␈α
suggestions␈α
of␈α
how␈α
to␈α
proceed:
␈↓ ↓H

␈↓ ↓H
␈↓ α_1.␈α∃The␈α∃information␈α∃that␈α∃I␈α∃have␈α∃seen,␈α∃including␈α∀last␈α∀years␈α∀budget,␈α∀do␈α∀not␈α∀contain␈α∀the
␈↓ ↓H
␈↓ ↓Hinformation␈αrequired␈αto␈αform␈αan␈αopinion␈αon␈αwhat␈αthe␈αadministrative␈αexpenses␈αbuy␈αus.␈αIn␈αparticular,
␈↓ ↓H
␈↓ ↓Hthe␈α
increase␈α
beyond␈α
inflation␈α
in␈α
these␈α
expenses␈α
of␈α
$572,000␈α
seems␈α
inadequately␈α
justified.
␈↓ ↓H

␈↓ ↓H
␈↓ α_2.␈α∪Professor␈α∪Massy's␈α∪article␈α∪in␈α∪␈↓↓Campus␈α∪Report␈↓␈α∪describes␈α∪a␈α∪very␈α∩sophisticated␈α∩method␈α∩of
␈↓ ↓H
␈↓ ↓Hanalysis,␈α
but␈α
the␈αexamples␈αhe␈αcites␈αof␈αrecent␈αadministrative␈αincreases␈α-␈αan␈αincrease␈αin␈αthe␈αPersonnel
␈↓ ↓H
␈↓ ↓HDepartment␈αand␈αa␈α
computer␈α
system␈α
to␈α
inform␈α
faculty␈α
of␈α
research␈α
opportunities␈α
both␈α
seem␈α
dubious.␈α
If
␈↓ ↓H
␈↓ ↓Hthe␈α∂latter␈α∂involves␈α∂even␈α∂one␈α∂salary␈α∂or␈α∂the␈α∂equivalent␈α∂in␈α∂computer␈α∂costs,␈α∂it␈α∞is␈α∞almost␈α∞certainly␈α∞not
␈↓ ↓H
␈↓ ↓Hworth␈α
it.␈α
If␈α
it␈α
isn't␈α
that␈α
expensive,␈α
it␈αmay␈αbe␈αworth␈αthe␈αcost,␈αbut␈αits␈αinclusion␈αin␈αhis␈αessay␈αis␈αthen␈αa
␈↓ ↓H
␈↓ ↓Hred␈α
herring.
␈↓ ↓H

␈↓ ↓H
␈↓ α_3.␈α∂There␈α∂are␈α∂three␈α∂reasons␈α∂why␈α∂administrative␈α∂expenses␈α∂are␈α∂not␈α∂subjected␈α∂to␈α∂as␈α∂careful␈α∞an
␈↓ ↓H
␈↓ ↓Hexamination␈α⊃as␈α⊂academic␈α⊂expenses.␈α⊂First,␈α⊂the␈α⊂effects␈α⊂of␈α⊂not␈α⊂filling␈α⊂a␈α⊂faculty␈α⊂slot␈α⊂are␈α⊂known␈α⊂and
␈↓ ↓H
␈↓ ↓Hlimited,␈α
whereas␈α
it␈α
can␈α
be␈αclaimed␈αthat␈αgetting␈αthe␈αpayroll␈αout␈α␈↓↓absolutely␈αrequires␈↓␈αanother␈αemployee.
␈↓ ↓H
␈↓ ↓HThe␈α∞expansion␈α∞of␈α∞the␈α∞personnel␈α∞department␈α∞Massy␈α∞refers␈α∞to␈α∞is␈α∞much␈α∞harder␈α
to␈α
evaluate␈α
than␈α
the
␈↓ ↓H
␈↓ ↓Hconsequences␈αof␈αletting␈αan␈αassociate␈αprofessorship␈αin␈αcomputer␈αscience␈αlapse.␈αSecond,␈αthe␈αofficials␈αof
␈↓ ↓H
␈↓ ↓Hthe␈α∂University␈α∂themselves␈α∂set␈α∂in␈α∂motion␈α∞administrative␈α∞projects␈α∞whose␈α∞mere␈α∞continuation␈α∞requires
␈↓ ↓H
␈↓ ↓Hfurther␈α∞growth,␈α∞and␈α∞they␈α∞are␈α∞reluctant␈α∞to␈α∞kill␈α∞their␈α∞own␈α
brain␈α
children.␈α
Third,␈α
the␈α
response␈α
to␈α
all
␈↓ ↓H
␈↓ ↓Hsorts␈α
of␈α
social␈α
demands␈α
has␈α
taken␈α
the␈α
form␈α
of␈α
expansion;␈α
the␈α
example␈α
I␈α
have␈α
in␈α
mind␈α
might␈αgive
␈↓ ↓H
␈↓ ↓Hoffense.
␈↓ ↓H

␈↓ ↓H

␈↓ ↓H

␈↓ ↓H

␈↓ ↓H

␈↓ ↓H
␈↓ α_4.␈α⊃The␈α⊂area␈α⊂wherein␈α⊂I␈α⊂have␈α⊂specific␈α⊂information␈α⊂is␈α⊂computing.␈α⊂In␈α⊂1971␈α⊂I␈α⊂wrote␈α⊂President
␈↓ ↓H
␈↓ ↓HLyman␈α⊃a␈α⊃letter,␈α⊃which␈α⊃he␈α⊃didn't␈α⊃acknowledge,␈α⊃complaining␈α⊂that␈α⊂Stanford␈α⊂spent␈α⊂more␈α⊂money␈α⊂on
␈↓ ↓H
␈↓ ↓Hadministrative␈αcomputing␈αthan␈αany␈αother␈αuniversity␈αin␈αproportion␈αto␈α
operating␈α
budget␈α
or␈α
number␈α
of
␈↓ ↓H
␈↓ ↓Hstudents.␈α∩Since␈α∩then,␈α∩the␈α∩sitution␈α⊃has␈α⊃gotten␈α⊃much␈α⊃worse.␈α⊃Stanford␈α⊃has␈α⊃purchased␈α⊃a␈α⊃completely
␈↓ ↓H
␈↓ ↓Hunnecessary␈α∞two␈α∞and␈α∞a␈α∞half␈α
million␈α
dollar␈α
computer␈α
that␈α
does␈α
only␈α
administrative␈α
work,␈α
the␈α
SCIP
␈↓ ↓H
␈↓ ↓Hstaff␈αhas␈αgrown␈αto␈α300,␈α
the␈α
merger␈α
of␈α
computing␈α
activities␈α
has␈α
led␈α
to␈α
academic␈α
computing␈α
supporting
␈↓ ↓H
␈↓ ↓Hadministrative,␈α∂and␈α∂an␈α∂unrealistic␈α∂amortization␈α∂decision␈α∂has␈α∂concealed␈α∂the␈α∂size␈α∂of␈α∂these␈α∂expenses.
␈↓ ↓H
␈↓ ↓HPart␈αof␈αthe␈αproblem␈α
is␈α
that␈α
the␈α
Provost␈α
and␈α
Vice-Provost␈α
for␈α
research␈α
are␈α
personally␈α
involved␈α
in␈α
the
␈↓ ↓H
␈↓ ↓Hnew␈αcomputer␈αscheme␈αwhose␈αexpenses␈αhave␈αgrown␈αand␈αgrown␈αand␈αseem␈αincapable␈αof␈αevaluating␈α
the
␈↓ ↓H
␈↓ ↓Hmonster␈α
they␈α
have␈α
created.
␈↓ ↓H

␈↓ ↓H
␈↓ α_It␈α∞will␈α∞be␈α∞very␈α∞difficult␈α∞to␈α∞evaluate␈α∞the␈α∞cost-effectiveness␈α∞of␈α∞the␈α∞administration.␈α∞Therefore,␈α
I
␈↓ ↓H
␈↓ ↓Hsuggest␈α
the␈α
following:
␈↓ ↓H

␈↓ ↓H
␈↓ α_1.␈α∀The␈α∀Budget␈α∀Priorities␈α∀Advisory␈α∀Commission␈α∀should␈α∪acquire␈α∪a␈α∪staff␈α∪of␈α∪two␈α∪or␈α∪three
␈↓ ↓H
␈↓ ↓Hincluding␈α⊂accounting␈α∂capability␈α∂to␈α∂look␈α∂into␈α∂the␈α∂possibility␈α∂of␈α∂cuts␈α∂in␈α∂administrative␈α∂expenses.␈α∂In
␈↓ ↓H
␈↓ ↓Horder␈αto␈αavoid␈αconflict␈αof␈αinterest␈αit␈α
should␈α
not␈α
be␈α
under␈α
the␈α
control␈α
of␈α
the␈α
ex-officio␈α
members␈α
of␈α
the
␈↓ ↓H
␈↓ ↓HCommission␈α
from␈α
the␈α
Administration.␈α
This␈α
staff␈α
should␈α
take␈α
an␈α
adversary␈α
position␈α
with␈α
regard␈αto
␈↓ ↓H
␈↓ ↓Hthe␈α
administrative␈α
expenditures.␈α
If␈α
this␈α
is␈α
too␈α
difficult,␈α
then␈α
the␈α
Commission␈α
should␈α
be␈αdivided␈αor
␈↓ ↓H
␈↓ ↓Hotherwise␈α_reorganized.␈α_There␈α_should␈α_be␈α_a␈α_report␈α_that␈α_breaks␈α↔down␈α↔personnel␈α↔and␈α↔money
␈↓ ↓H
␈↓ ↓Hexpenditures␈α∞with␈α∞regard␈α∞to␈α∞function,␈α
and␈α
the␈α
consequences␈α
of␈α
omitting␈α
or␈α
reducing␈α
each␈α
function
␈↓ ↓H
␈↓ ↓Hshould␈α
be␈α
spelled␈α
out.
␈↓ ↓H

␈↓ ↓H
␈↓ α_I␈α∞must␈α∞say␈α∞that␈α∞the␈α∞Budget␈α∞Priorities␈α∞Advisory␈α∞Commission␈α∞seems␈α∞to␈α
have␈α
entirely␈α
followed
␈↓ ↓H
␈↓ ↓Hthe␈αline␈αof␈αinvestigation␈αsuggested␈αby␈αthe␈αAdministration␈αand␈αdoesn't␈αseem␈αto␈αbe␈αgiving␈αany␈αweight
␈↓ ↓H
␈↓ ↓Hto␈α
the␈α
idea␈α
that␈α
major␈α
economies␈α
in␈α
administration␈α
are␈α
possible.
␈↓ ↓H

␈↓ ↓H
␈↓ α_2.␈α∩A␈α∩formal␈α∩comparison␈α∩of␈α⊃each␈α⊃major␈α⊃function␈α⊃between␈α⊃Stanford␈α⊃and␈α⊃other␈α⊃universities
␈↓ ↓H
␈↓ ↓Hshould␈α∂be␈α∂made␈α∂in␈α∂the␈α∂hopes␈α∂of␈α∂finding␈α∂another␈α∂university␈α∂that␈α∂accomplishes␈α∂the␈α∂function␈α∂more
␈↓ ↓H
␈↓ ↓Hcheaply␈αthan␈αStanford␈αdoes.␈αOtherwise,␈αit␈αmay␈αbe␈αtoo␈αeasy␈αto␈αaccept␈αpredictions␈αthat␈αdisaster␈αwould
␈↓ ↓H
␈↓ ↓Hfollow␈α⊗any␈α⊗cut.␈α⊗Perhaps␈α⊗the␈α⊗comparisons␈α⊗should␈α⊗become␈α⊗available␈α⊗only␈α⊗after␈α∃a␈α∃preliminary
␈↓ ↓H
␈↓ ↓Hevaluation␈α
of␈α
the␈α
consequences␈αof␈αcuts␈αis␈αmade␈αso␈αthat␈αthe␈αcomparison␈αcan␈αbe␈αused␈αto␈αcalibrate␈αthe
␈↓ ↓H
␈↓ ↓Hevaluation.␈α∩A␈α∩preliminary␈α∩overall␈α∩comparison␈α∩could␈α∩probably␈α⊃be␈α⊃made␈α⊃rather␈α⊃quickly␈α⊃with␈α⊃say
␈↓ ↓H
␈↓ ↓HHarvard,␈α
M.I.T.,␈α
Yale,␈α
Princeton,␈α
and␈α
Caltech.
␈↓ ↓H

␈↓ ↓H
␈↓ α_3.␈α
In␈α
the␈α
particular␈α
area␈α
of␈α
computing,␈α
I␈α
can␈αbe␈αmore␈αconcrete.␈αThe␈αstaff␈αof␈αSCIP␈αshould␈αbe
␈↓ ↓H
␈↓ ↓Hcut␈α∞in␈α∞half␈α∞and␈α∞the␈α∞IBM␈α∞370/158␈α∞should␈α
be␈α
sold.␈α
Academic␈α
computing␈α
should␈α
again␈α
be␈α
separated
␈↓ ↓H
␈↓ ↓Hfrom␈α⊃administrative␈α⊃computing,␈α⊂and␈α⊂the␈α⊂academic␈α⊂computing␈α⊂center␈α⊂should␈α⊂be␈α⊂run␈α⊂by␈α⊂a␈α⊂faculty
␈↓ ↓H
␈↓ ↓Hcommittee.␈α∞As␈α
many␈α
as␈α
possible␈α
of␈α
the␈α
administrative␈α
computing␈α
functions␈α
should␈α
be␈α
contracted␈α
to
␈↓ ↓H
␈↓ ↓Hspecialist␈α
firms;␈α
included␈α
in␈α
this␈α
should␈α
certainly␈α
be␈α
payroll␈α
and␈α
medical␈α
billing␈α
both␈α
of␈αwhich␈αare
␈↓ ↓H
␈↓ ↓Hdone␈α
more␈α
cheaply␈α
on␈α
the␈α
outside.
␈↓ ↓H

␈↓ ↓H
␈↓ α_After␈α⊃writing␈α⊃the␈α⊃above,␈α⊃I␈α⊃must␈α⊃confess␈α⊃that␈α⊃I␈α⊃am␈α⊃not␈α⊃entirely␈α⊃confident␈α⊂of␈α⊂all␈α⊂the␈α⊂facts.
␈↓ ↓H
␈↓ ↓HHowever,␈α∞I␈α∞have␈α∞discussed␈α∞the␈α∞problem␈α∞with␈α∞many␈α∞other␈α∞faculty␈α
members,␈α
and␈α
I␈α
have␈α
discovered
␈↓ ↓H
␈↓ ↓Hthat␈α∂they␈α∂know␈α∂even␈α∂less,␈α∂and␈α∂I␈α∂have␈α∂discovered␈α∂that␈α∂the␈α∂high-level␈α∂University␈α∂officials␈α∂and␈α∞the
␈↓ ↓H
␈↓ ↓Hofficials␈αof␈αSCIP␈αaren't␈αeven␈αasking␈αthe␈αright␈αquestions.␈αSo␈αI␈αfeel␈αjustified␈αin␈αpressing␈αthe␈αmatter␈αin
␈↓ ↓H
␈↓ ↓Hspite␈α
of␈α
my␈α
lack␈α
of␈α
knowledge.
␈↓ ↓H

␈↓ εW
Sincerely yours,
␈↓ εW

␈↓ εW

␈↓ εW

␈↓ εW

␈↓ εW

␈↓ εW
John McCarthy
␈↓ εW
Director, Artificial Intelligence Laboratory
␈↓ εW
Professor of Computer Science
␈↓ εW

␈↓ ↓H
JMC:pw
␈↓ ↓H
juilla.le1[let,jmc]:SU-AI
␈↓ ↓H