MAY 7 1975
Dear Sauls AY 7 ID

We Rave been giving a good deal of thought and attention
to the issves raised by your Propesal to augment your 10 at Rutgers,
This response has been discussed with Ed Feigenbaum and Fllfet Levinthal
ee well as with Tomy bue IT felt we should not carry it beyend oyr own
two groups without further eonsultation with you,

Briefly, there are two maim reasons why I do mot think that we
should recommend the use of the SUMEX@AIM grant as a vehicle for this
Puroose,

1) T believe that this pattern of extension of the SUMEX=

AIM system is a significant departure from the charter that was aythore
ized by the RR Council ew imvolving as it dees not onty a siqnificant
capital investment, but am INEVITABLE commitment for further operational
BuppoFt In the future

and 2) serious technical and especially MANAGERIAL problems
of supporting two distant systems im a way that would really accemplish
the backup goals that are the rationale in the present proposal,

In the light of our extensive experience with efforts te sustain
day to day compatibility with IMSSS and with other network TENEXes, our
Staff is Umanmimous that only an authoritative single management of the
two facilities would make it actually work to these goals, I am syre
that you would be as loathe to delegate such responsibility as we would
be to assume its at its best, and even with "*ul!l authority", fe would
be a mamagerial mightmare,

We do have an alternative proposal that would answer the
technical issues of 2), Amd as for 1), we would afl) offer yoy the
strongest support im defense of a project propesal from you before the
appropriate study seetion and Coumeil, Amd we can transmit as much
more detai! as you wish about the technical problems w= the simplest
would be for vou to talk te Tom and Rainer,

The alternative option that we have in mimd would be an augmene
tation of your system with the aim of giving the SUMEX#AIM community
access to TOPSe10 as a complement to TENEX,

This would greatiy bemefit SUMEX=AIM ysers who are having
difficulties im adapting their programs to TENEX{! We believe many of these
problems will eventually be solved, but some perhaps on)y at unreasonable ex
Pemse compared to Just having time available on a TOPS#10 machine, Also the
conversions would be very much facilitated i# we could examine the punning
eode as {t was IN FACT being executed under TOPS#10,

In our opinions, thia widening of options te SUMEX@AIM users would be
of greater benefit as a return for augmenting your system, tham would be
the exact duplication of SUMEX=tenex, With efficient communication between
eure two systems, there may even be ocecasions when users would Pun Progpam
segments aiternately under TENEX and under TOPS#10 in accordance with the
availability of the appropriate software,In fact there are a number of
very imteresting experiments om interesystem coordination amd coopera»
tiom that would be advanced by this complementarity, and wil! be of ute
most importance in the future development of compyter networking and
the concept of division of Jabor among functionally specialized centers,
tr the Om Lernoylre PR BOY - Qian
fractionation of capital investmente= and commitment to the

future costs that will follow with {¢ == are mot the most
cost effective use of available funds, (For example, im a year or two, when
we are more heavily loaded, the upgrading of the KI to a KL processor can
Probably give us a 2 or 3X imerease im thruput at relatively little cost, We
are at a poimt om the investmentayield curve (taking stating also inte
account) where it makes jittie sense to distribute the investment into frag=
MOntS,

We can sympathize that ome motive for lecal avamentation may be
the stil! imperfect state of communications across the country compared
to the local phone cali, We think this is subject to continued improvement
and certaimiy wil) support any reasonable steps to make working on the
System more comfortable for you and your colleagues,

Josh
Dear Sauls

We Rave been giving a good deal of thought and attention
to the {ssues raised by your proposal to augment your 10 at Rutgers,
This response has been discussed with Ed Feigenbaum and Eltfot Levinthal
as well as with Tomy but I felt we should not carry it beyond our own
two groups without further eonsultation with you,

Brieflys there are two main reasons why I do not think that we
should recommend the use of the SUMEX*AIM grant as a vehicle for this
puroose,

4) I belleve that this pattern of extension af the SUMEX«

AIM system is a significant departure from the charter that was authors
ized by the RR Couned! == involving as {t does net anly a significant
eabita! investments but am INEVITABLE commitment for further eperationa)
Support tm the future

and 2) serious technical and especially MANAGERIAL problems
of supporting two distant systema jin a way that would really accomplish
the backup goals that are the rationale in the present propesal,

Im the light of our extensive experience with efforts ta sustain
day to day compatibility with IMSSS and with ether metwork TENEXes, our
staff is unanimous that only am authoritative single management of the
two facilities would make it actually work to those goals, I am sure
that you would be as loathe to delegate such responsibility as we would
be to assume {tt at its beet, and even with "full authority", ft would
be a managerial nightmare,

We do have an alternative proposal that would answer the
technical issues of 2), And as for 1), we would al! offer you the
strongest support in defense of a project proposal from you before the
appropriate study section and Counei}, And we can transmit as much
more detail es you wish about the technica! problems = the simplest
would be for you to talk ¢o Tom and Rainer,

The alternative option thet we have in mind would be am augmene
tation of your system with the aim of giving the SUMEX#AIM community
access te TOPS#10 as a complement to TENEX,

ThiS would greatly benefit SUMEXAIM users whe are having
diff(eulties im adapting thelr programs to TENEX} We believe many of these
Problems wil! eventually be solved, but some perhaps only at unreasonable exe
pense compafed to just having time available on a TOPS#10 machine, Also the
conversions would be very much facilitated if we could examine the running
code as it was IN FACT being executed under TOPS#10,

Im our opiniom, this widening of options te SUMEX@AIM Users would be
of greater benefit as a return for augmenting your system, than would be
the exact duplication of SUMEXetenex, With efffeient communication between
our two Systems, there may even be occasions when users would rum pregram
Segments alternately under TENEX and under TOPS#10 im accordance with the
aveilabijity of the appropriate software,In fact there are a number of
Very interesting experiments om inter=system coordination and coopera
tiom that would be advanced by this complementarity, and wil!) be of ute»
most {mportamce {nm the future development of computer networking and
the concept of division of labor among functionally specialized centers,
t, the
fractionation of capital imvestment== and commitment to the

future costs that will follow with it -=# are not the most

cost effective use of available funds, [For example, im a year or two, when
we are more heavily loaded, the upgrading of the KI to a KL processor can
PPobebly give us a 2 of 3X imerease in thruput at relatively little eost, We
are at a point on the investmemtaylield curve (taking staffing also into
account) where {t makes little sense to distribute the investment inte frage
ments,

_ We cam sympathize that one motive for toeal augmentation may be
the still imperfect state of communications acress the country compared
to the local phone call, We think this {s subleet to continued improvement
and certainly will support any reasonable steps to make working on the
System more comfortable for you and your colleagues,

Josh