/FONT#0=BASL30/FONT#1=BASI30/FONT#2=NGR40/FONT#3=NGR25/FONT#4=NGR20/FONT#15=STA200
␈↓␈↓↓␈↓α␈↓β␈↓∧␈↓∂␈β↓↓␈↓∂S␈↓α
␈↓ ↓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
␈↓␈↓ ¬l24 October 1974
␈↓ ↓H

␈↓ ↓H

␈↓ ↓H

␈↓ ↓H

␈↓ ↓H

␈↓ ↓H

␈↓ ↓H

␈↓ ↓H

␈↓ ↓H

␈↓ ↓H
Richard Atkinson
␈↓ ↓H
Associate Dean
␈↓ ↓H
Humanities and Science
␈↓ ↓H
Stanford University
␈↓ ↓H
Stanford, California 94305
␈↓ ↓H

␈↓ ↓H
Dear Dick:
␈↓ ↓H

␈↓ ↓H
␈↓ α_Enclosed␈αis␈αthe␈αcopy␈αyou␈αrequested␈αof␈αmy␈α1971␈αletter␈αto␈αLyman.␈αFrom␈αwhat␈αyou␈αtold␈αme␈αtoday
␈↓ ↓H
␈↓ ↓Habout␈αthe␈αpotential␈αimpact␈αof␈αthe␈αbudget␈αcuts␈αon␈αthe␈αinstructional␈αprogram,␈αthe␈αissues␈αraised␈αin␈αthat
␈↓ ↓H
␈↓ ↓Hletter␈α
are␈α
much␈α
more␈α
urgent␈α
now␈α
than␈α
they␈α
were␈α
at␈α
that␈α
time.
␈↓ ↓H

␈↓ ↓H
␈↓ α_The␈α∞state␈α∞of␈α∞my␈α∞information␈α∞(or␈α∞anyone␈α∞else's)␈α∞about␈α∞the␈α∞costs␈α∞of␈α∞computing␈α∞is␈α∞much␈α
worse
␈↓ ↓H
␈↓ ↓Hnow␈α∞than␈α∞it␈α∞was␈α∞then.␈α∞At␈α
that␈α
time,␈α
administrative␈α
computing␈α
was␈α
a␈α
separate␈α
organization␈α
and␈α
its
␈↓ ↓H
␈↓ ↓Hcosts␈αcould␈αbe␈αestimated.␈αThe␈αmerger␈αof␈αall␈αcomputing␈αactivities␈α
into␈α
SCIP␈α
has␈α
made␈α
it␈α
very␈α
difficult
␈↓ ↓H
␈↓ ↓Hto␈α∞allocate␈α∞costs␈α∞properly.␈α∞However,␈α∞I␈α∞think␈α∞that␈α∞the␈α∞potential␈α∞savings␈α∞from␈α∞reorganizing␈α∞Stanford
␈↓ ↓H
␈↓ ↓Hcomputing␈α
are␈α
much␈α
larger␈α
now␈α
than␈α
they␈α
were␈α
at␈α
that␈α
time.
␈↓ ↓H

␈↓ ↓H
␈↓ α_My␈α
guesstimate␈α
of␈α
the␈α
potential␈α
savings␈α
is␈α
as␈α
follows:
␈↓ ↓H

␈↓ ↓H
␈↓ α_1.␈α∞If␈α
the␈α
360/67␈α
is␈α
considered␈α
fully␈α
amortized,␈α
its␈α
rates␈α
can␈α
be␈α
considerably␈α
reduced,␈α
and␈α
the
␈↓ ↓H
␈↓ ↓HProvost's␈αfund␈α
subsidy␈α
to␈α
academic␈α
computing␈α
can␈α
be␈α
cut␈α
in␈α
half␈α
with␈α
no␈α
reduction␈α
in␈α
the␈α
amount␈α
of
␈↓ ↓H
␈↓ ↓Hcomputing␈α≤available␈α≤to␈α≠students␈α≠and␈α≠faculty.␈α≠In␈α≠fact,␈α≠except␈α≠for␈α≠government␈α≠accounting
␈↓ ↓H
␈↓ ↓Hconsiderations,␈α⊂the␈α⊂fund␈α⊂could␈α⊂probably␈α⊂be␈α⊂completely␈α⊂eliminated.␈α⊂A␈α⊂renewed␈α⊂attempt␈α⊂to␈α⊂get␈α⊂the
␈↓ ↓H
␈↓ ↓Hgovernment␈α⊂accountants␈α⊂to␈α⊂look␈α⊂at␈α⊂the␈α⊂matter␈α⊂sympathetically␈α⊂would␈α⊂be␈α⊂justified␈α⊂in␈α⊂view␈α∂of␈α∂the
␈↓ ↓H
␈↓ ↓Hfinancial␈α
state␈α
of␈α
the␈α
university.␈α
This␈α
cost␈α
reduction␈α
would␈α
require
␈↓ ↓H

␈↓ ↓H
␈↓ αha.␈αNo␈αlonger␈α
charging␈α
academic␈α
computing␈α
with␈α
the␈α
costs␈α
of␈α
extending␈α
WYLBUR␈α
etc.␈α
to
␈↓ ↓H
␈↓ ↓Hthe␈α370.␈αAt␈αpresent␈αthese␈αcosts␈αare␈αconsidered␈αgeneral␈αcosts␈αof␈αSCIP␈αand␈αare␈αallocated␈αin␈αproportion
␈↓ ↓H
␈↓ ↓Hto␈α∞revenues.␈α∞At␈α∞present␈α∞the␈α∞67␈α∞accounts␈α∞for␈α∞$2.6␈α∞million␈α∞out␈α∞of␈α∞$4.1␈α∞million␈α
in␈α
revenues␈α
and␈α
so␈α
is
␈↓ ↓H
␈↓ ↓Hcharged␈α70␈αpercent␈αof␈αan␈α
expense␈α
that␈α
leads␈α
to␈α
little␈α
if␈α
any␈α
improvement␈α
in␈α
the␈α
facilities␈α
available␈α
on
␈↓ ↓H
␈↓ ↓Hthe␈α∂67.␈α∂It␈α∂will␈α∂also␈α∞be␈α∞necessary␈α∞to␈α∞reduce␈α∞the␈α∞free␈α∞courses␈α∞and␈α∞the␈α∞consulting.␈α∞␈↓↓A␈α∞target␈α∞personnel
␈↓ ↓H
␈↓ ↓Hfigure␈αshould␈αbe␈αthe␈αnumber␈αof␈αpeople␈αemployed␈αwhen␈αthe␈αIBM␈α7090␈αwas␈αthe␈αmain␈αcampus␈αcomputer␈↓.
␈↓ ↓H
␈↓ ↓HThe␈α
estimate␈α
of␈α
possible␈α
cost␈α
reduction␈α
is␈α
based␈α
on␈α
the␈α
fact␈α
the␈α
67␈α
is␈α
paid␈α
for.
␈↓ ↓H

␈↓ ↓H

␈↓ ↓H

␈↓ ↓H

␈↓ ↓H

␈↓ ↓H

␈↓ ↓H
␈↓ α_2.␈αJudging␈αfrom␈αwhat␈αother␈αuniversities␈αwere␈αpaying␈αin␈α1971,␈αI␈αthink␈αthe␈αbusiness␈αcomputing
␈↓ ↓H
␈↓ ↓Hof␈α⊃the␈α⊃university␈α⊃can␈α⊃be␈α⊃done␈α⊃for␈α⊃$500,000␈α⊃per␈α⊃year.␈α⊃Serious␈α⊃consideration␈α⊃should␈α⊃be␈α⊃given␈α⊂to
␈↓ ↓H
␈↓ ↓Hcontracting␈α∞all␈α∞or␈α∞most␈α∞of␈α∞it␈α∞to␈α∞outside␈α∞organizations,␈α∞e.g.␈α
Bank␈α
of␈α
America␈α
should␈α
be␈α
asked␈α
for␈α
a
␈↓ ↓H
␈↓ ↓Hquote␈α
on␈α
the␈α
payroll.
␈↓ ↓H

␈↓ ↓H
␈↓ α_Here␈α
are␈α
some␈α
random␈α
facts␈α
in␈α
support␈α
of␈α
the␈α
idea␈α
that␈α
large␈α
savings␈α
are␈α
to␈α
be␈α
made:
␈↓ ↓H

␈↓ ↓H
␈↓ α_1.␈αSCIP␈αhas␈α300␈αemployees,␈αi.e.␈αis␈αone␈αthird␈αthe␈αsize␈αof␈α
the␈α
faculty.␈α
The␈α
growth␈α
of␈α
this␈α
number
␈↓ ↓H
␈↓ ↓Hwith␈α
time␈α
should␈α
be␈α
examined.
␈↓ ↓H

␈↓ ↓H
␈↓ α_2.␈α
Some␈α
years␈α
ago,␈α
SCIP␈α
quoted␈α
the␈α
psychiatric␈αclinic␈αa␈αprice␈αof␈α$90,000␈αper␈αyear␈αfor␈αpatient
␈↓ ↓H
␈↓ ↓Hbilling␈α⊗and␈α⊗the␈α⊗clinic␈α⊗got␈α⊗an␈α⊗outside␈α⊗quote␈α⊗of␈α⊗$15,000.␈α∃I␈α∃believe␈α∃the␈α∃matter␈α∃was␈α∃somehow
␈↓ ↓H
␈↓ ↓Hcompromised␈αand␈αSCIP's␈αmonopoly␈αwas␈αpreserved,␈αbut␈αI␈αsuppose␈αthis␈αwas␈αdone␈αby␈αputting␈αmore␈αof
␈↓ ↓H
␈↓ ↓Hthe␈α
burden␈α
on␈α
an␈α
organization␈α
that␈α
could␈α
not␈α
make␈α
a␈α
comparison␈α
of␈α
outside␈α
costs.
␈↓ ↓H

␈↓ ↓H
␈↓ α_3.␈α
SCIP␈α
quoted␈α
$21,000␈α
per␈α
semester␈α
to␈α
run␈α
CS206␈α
on␈α
an␈αinteractive␈αbasis␈αnot␈αincluding␈αthe
␈↓ ↓H
␈↓ ↓Hcost␈α
of␈α
the␈α
terminals␈α
to␈α
be␈α
used.␈α
IMSS␈α
quoted␈α
$4500␈α
for␈α
the␈α
same␈α
service␈αincluding␈αthe␈αuse␈αof␈αthe
␈↓ ↓H
␈↓ ↓Hterminals.
␈↓ ↓H

␈↓ ↓H
␈↓ α_4.␈αA␈α(chemistry?)␈αprofessor␈αsurveyed␈αcosts␈αof␈αlarge␈α
scale␈α
computing␈α
and␈α
discovered␈α
he␈α
could␈α
do
␈↓ ↓H
␈↓ ↓Hit␈α
for␈α
1/10␈α
the␈α
price␈α
outside␈α
Stanford.
␈↓ ↓H

␈↓ ↓H
␈↓ α_There␈α
are␈α
many␈α
more␈α
examples␈α
of␈α
this␈α
kind,␈α
but␈α
I'll␈α
stop,␈αbecause␈αthe␈αprobability␈αof␈αhaving
␈↓ ↓H
␈↓ ↓Hany␈α
influence␈α
is␈α
too␈α
low␈α
to␈α
justify␈α
more␈α
effort␈α
even␈α
to␈α
the␈α
extent␈α
of␈α
writing␈α
this␈α
letter␈α
more␈α
carefully.
␈↓ ↓H

␈↓ ↓H
␈↓ α_Here␈αis␈αa␈αconjecture␈αon␈αwhy␈αthis␈αgrowth␈αhas␈αoccurred,␈αbut␈αI␈αdon't␈αknow␈αthe␈αinternal␈αsituation
␈↓ ↓H
␈↓ ↓Hwell␈αenough␈αto␈αbe␈αsure␈αof␈αit.␈αSCIP␈αhas␈αundertaken␈αmany␈αprojects,␈αSPIRES,␈αBALLOTS,␈αWYLBUR,
␈↓ ↓H
␈↓ ↓HORVYL,␈α∞OASIS.␈α∞The␈α∞ideas␈α∞for␈α∞most␈α∞of␈α∞these␈α∞projects␈α∞were␈α∞generated␈α
internally,␈α
but␈α
most␈α
of␈α
the
␈↓ ↓H
␈↓ ↓Hcosts␈α∞have␈α∞at␈α∞least␈α∞nominally␈α∞come␈α∞from␈α∞the␈α∞outside.␈α∞I␈α∞say␈α∞nominally,␈α
because␈α
it␈α
is␈α
often␈α
arbitrary
␈↓ ↓H
␈↓ ↓Hwhether␈α
a␈αsystem␈αcost␈αthat␈αmakes␈αa␈αproject␈αpossible␈αshould␈αbe␈αcharged␈αto␈αthe␈αproject␈αor␈αconsidered
␈↓ ↓H
␈↓ ↓Hthe␈α∂cost␈α∂of␈α∂a␈α∂general␈α∂improvement.␈α∂An␈α∂organization␈α∂like␈α∂SCIP␈α∂that␈α∂is␈α∂not␈α∞subject␈α∞to␈α∞overall␈α∞cost
␈↓ ↓H
␈↓ ↓Hcontrol␈α∂tends␈α∂to␈α∂bid␈α∂low␈α∂to␈α∞a␈α∞potential␈α∞sponsor␈α∞of␈α∞a␈α∞new␈α∞activity.␈α∞An␈α∞extreme␈α∞example␈α∞is␈α∞OASIS
␈↓ ↓H
␈↓ ↓Hwhich␈αarose␈αthrough␈αa␈α$200,000␈α
Ford␈α
grant␈α
many␈α
years␈α
ago␈α
and␈α
whose␈α
development␈α
cost␈α
many␈α
times
␈↓ ↓H
␈↓ ↓Hthat␈α
before␈α
it␈α
was␈α
abandoned.
␈↓ ↓H

␈↓ ↓H
␈↓ α_The␈α∞amortization␈α
of␈α
the␈α
370/158␈α
over␈α
ten␈α
years␈α
is␈α
questionable␈α
and␈α
mortgages␈α
the␈α
future␈α
of
␈↓ ↓H
␈↓ ↓HStanford␈α
computing␈α
for␈α
a␈α
very␈α
long␈α
time.
␈↓ ↓H

␈↓ ↓H
␈↓ α_A␈αdetailed␈αcost␈αanalysis␈αwould␈αbe␈αneeded␈αto␈αdetermine␈α
the␈α
true␈α
allocation␈α
of␈α
costs␈α
and␈α
the␈α
very
␈↓ ↓H
␈↓ ↓Hbest␈αcourse␈αto␈αfollow.␈αHowever,␈αin␈αthis␈αtime␈αof␈αextreme␈αbudget␈αstringency,␈αa␈αcruder␈αapproach␈αshould
␈↓ ↓H
␈↓ ↓Halso␈α∞be␈α∞considered.␈α∞This␈α∞is␈α∞to␈α∞bring␈α∞in␈α∞a␈α∞new␈α∞manager␈α∞without␈α∞commitments␈α
to␈α
present␈α
personnel
␈↓ ↓H
␈↓ ↓Hand␈αrequire␈αof␈αhim␈αa␈αspecific␈αcut␈αin␈αpersonnel␈αan␈αoperating␈αcosts.␈αIt␈αwould␈αprobably␈αalso␈αbe␈αwise␈αto
␈↓ ↓H
␈↓ ↓Hseparate␈α∪academic␈α∩computing␈α∩from␈α∩administrative␈α∩and␈α∩ask␈α∩the␈α∩faculty␈α∩to␈α∩allocate␈α∩cuts␈α∩between
␈↓ ↓H
␈↓ ↓Hcomputing␈α
and␈α
other␈α
parts␈α
of␈α
the␈α
academic␈α
budget.
␈↓ ↓H

␈↓ ↓H

␈↓ ↓H

␈↓ ↓H

␈↓ ↓H

␈↓ ↓H
␈↓ α_This␈α∞letter␈α∞which␈α∞you␈α∞requested␈α∞is␈α∞absolutely␈α∞the␈α∞last␈α∞comment␈α∞I␈α∞shall␈α∞make␈α
on␈α
the␈α
costs␈α
of
␈↓ ↓H
␈↓ ↓HStanford␈αcomputing␈αunless␈αmore␈αis␈αsolicited␈αby␈αLyman␈αor␈αMiller.␈αI␈αhave␈αbeen␈αinterested␈αin␈αStanford
␈↓ ↓H
␈↓ ↓Hcomputing␈α∪for␈α∪a␈α∪number␈α∪of␈α∪years␈α∩and␈α∩have␈α∩taken␈α∩part␈α∩in␈α∩various␈α∩low␈α∩level␈α∩committees␈α∩with
␈↓ ↓H
␈↓ ↓Hnegligible␈αresult␈αfor␈αthe␈αwork␈αput␈αin.␈αMy␈αconclusion␈αis␈αthat␈αadvice␈αis␈α␈↓↓not␈αwanted␈↓.␈αMy␈αonly␈α
reason␈α
for
␈↓ ↓H
␈↓ ↓Hacceding␈α∞to␈α∞your␈α∞request␈α∞this␈α∞time␈α∞is␈α∞that␈α∞you␈α∞may␈α∞be␈α∞able␈α
to␈α
avert␈α
some␈α
cuts␈α
in␈α
Humanities␈α
and
␈↓ ↓H
␈↓ ↓HScience␈αby␈αpointing␈αto␈αone␈αof␈αthe␈αadministrative␈αareas␈αthat␈α
can␈α
be␈α
cut.␈α
I␈α
have␈α
totally␈α
given␈α
up␈α
trying
␈↓ ↓H
␈↓ ↓Hto␈α∞get␈α∞Stanford␈α∞to␈α∞make␈α∞interactive␈α∞computing␈α∞available␈α∞to␈α∞the␈α
campus␈α
generally␈α
or␈α
to␈α
get␈α
a␈α
good
␈↓ ↓H
␈↓ ↓Hdisplay␈α
system␈α
for␈α
the␈α
Computer␈α
Science␈α
Department.
␈↓ ↓H

␈↓ ↓H
␈↓ α_I␈α
hope␈α
this␈α
is␈α
of␈α
some␈α
use.
␈↓ ↓H

␈↓ ↓H

␈↓ εW
Sincerely yours,
␈↓ εW

␈↓ εW

␈↓ εW

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

␈↓ ↓H
JMC:pw
␈↓ ↓H
atkins.le3[let,jmc]:SU-AI
␈↓ ↓H

␈↓ ↓H