x x xK OX x x x KK Ox x x x 
* * 
x x 











XK RK 2k * x KK KKK x Kk XK 
x OK x x x x OkKK x *K KX 
\ KK x x OOK ok * x x FORIOKK x x 


OKO OOOO SOO IRB SS OR ARKO SISO SSSR RKC IKK 

















* x 
x NOTICE x 
x x 
x This document may cantain information covered x 
x or more » COFUPightsy and x 
x logure are Circulation of x 
x t is req ed ta holders af @ * 
x UNIX fram x 
x Western SMa x 
x rerraduce t armic * 
* ty with + x 
x x 
* All oather cirevlation or rerroduetion is x 
x erohioited. i x 
x * 
OOOO COO OOO OOOO OR OOO OK OO ORK qOK x OK 






HARVART MEE 





As rreviously ammoune 
a at the Harvard Univers 
—Jober io and endings early 






Learn United State 
at 2:00 FHM an Friday Oc- 
tober 3. 








start 
aflernoan af Sunday Qe 








The Frogram committee mambe 
(212-694-3501) and Lew Law (617-498 


re Tucker Taft (617-495-3353) Lou Katz 
RGO2Q7 6 







For rear] 
take the ai 


a flying ta Rostans the auicl 
ta the @irreart subway ¢ 










route toa Harvard Sauare is to 
ations take a KRlue Line train to 












Government C here char tae mo Line ta ft ben’ then ary to a Red 
Line to Harverd Sau + Thi kas 20 t rs The alternative is to 





teke a cah at @ cost of ahout $7.00. 


Accammodations: Sheraton Cand 






ar Hotels 16 Garden Street 

a4: ROO 

Holiday Inmy 1651 Mase. Avenue 
617-491-1000 

Harvard Motor Hot y 110 Mb. Auburn St. 
617-864 00 






All are within walking distance of the Science 
the flir side. 





henter. A mar of the area is orn 


UNIX NEWS 


The Sertember issue will arrear after the Harvard meeting and will contain 
the membershire List and. Hotl infarmations The closing date for receirt of 
qrrections to the mailing list is October Sy 1976. 


Jyoat 






To date we have received checks from 374 af the installations. 





Prof, Melvin Ferentz Brookwln Callede of CUNY Brooklunys NY 11210 






armas 
calm tw 


eee 
SS Sie A 






x 






Hunoines & Geounos 
SEAVICE BULDINGS, 


SN 

SA 
a “w 

crake, 


Mata yoy Huey 














' ee NE SN 
J SOLDIER'S FIELD ty x 
ate 
i p 
A O a a 


. 


WON TY 
StUbiVs 


AUSINESS. SCHOOL 
PAR ARIES 
vor 





3 
om 
se 






otis 
iri 


ry 


10 Scmoots 


: ' 
- a{em tasiny 


iMG Lor 


wr pIciNE, 


DENTISTRY, CUA en Ag TH, 
ARK OLD ARRCAT TUM FTE 














Copyright 1972 by the 
Presutent and Fellows of Harvard 





aQwitta naayaA awpaaIna 
THE HEBREW UNIVERSITY OF JERUSALEM 


COMPUTER SCIENCE LABORATORY 
INSTITUTE OF MATHEMATICS 


August 17, i976 


Professor M. Perentz 
Physics Department 
Brooklyn College of CUNY 
Brooklyn, New York 11218 
U.S.A. 


Dear Professor Ferentz: 

After a long wait we finally have a working version of 
Unix. That is, for almost four months now we have been working 
under Unix, but single user and being extremly careful, bevause 
of our little BUG, which was described in our June 14th letter. 
But since yesterday, when an 8k bank of memory stopped working, 
and was taken out for repair, with it also went our timing bug. 
t will be guite interesting te see what happens when the bank 
is sgain operational. At least our Dee man now knows that it was 
a hardware problem ali along and Unix is as good as its fame. 


Our lab has: 


1 PDP 11/45 with no floating point 
48k “memory (8k MOS + 46 vore} 

i RKOS ( + one on its way) 

3 DLil 

1 Duli 

i vTil 

1 VARIAN electrostatic printer/plotter 
3 cDC 714 terminal 


We also have a NOVA 888 with 16k, which among its 
peripherals there is a Diablo Hytipe. 


We have so far developed software for file transfer to the 
NOVA and the CDC Ciber (this one belonging to the Computer Center) 
which work on an asynchronous interface (DL1i), but are working 
hard to connect the CDC through a synchronous interface (DU11). 
Noel Kropf is responsible for the driver of the VT11 and most 
debugging of the system, and since he has left for more fertile 
grounds, (i.e Columbia Univ.), we’ve lost a great help. 


Our main fields will be graphics, communications and cvompi- 
lers so anybody willing to share information on these is welvomed. 


Finally I want te thank David Barach at Yale Univ. for making 
us feel that we weren't alone in our struggle. 


2 Sincerely 


RRO 
~Daniek- aeadioe> 


watt) b> 


Molcolm Blanchord 24 August 1976 
Computer Grephics Lob 

New York Institute of Technology 

P.O. Box 178 

Old Westbury, New York 11568 


Prof. Melvin Ferentz.. 

Physics Deportment 

Brooklyn College of CUNY ; 
Brook:yn, Neu York [1218 % 


Bear Hel; 


I have discovered that [cneck with the -s switch. does not work 
for devices with more than 32768 blocks. The problem ts In 
icheck’s subroutine makefree. It wos treating the number of 
blocks as ao signed number. The following is @ fixed version of 
makefree. Be ouore that the printer that produced this listing 
prinis an underbar os “¢”. 


mokefree() 


t oe. 





register char *1z 
\\ register j: 
sblock.s¢nfree = Bz 
sblock.s¢ninode = 6: 
sblock.s¢fiock © 83 
sblock.s¢ilock = 8: 
sblock.s¢fmod = 83 
free(@): 
for(i=sblock.s¢fsize-l) id=sblock.s¢ 






= are 





i Tteropt 40877771 & CIKE 
freecid: 
+ 
burite(l, &sblock?: 
close(fids 
syncs 
return: 
2 
Sincerely, 


Nedwt. Bhob A 


Malcolm Blanchard = 


Jeez >=-8> May 





- Tor § 


Septenber 10, 1976 


Professor Melvin Ferentz , 
Physics Department , 
Brooklyn College of CUNY 
Brooklyn, New York 11210 


SUBJECT: “UNIX News" 
Dear Professcr Ferentz: 


We have discovered a bug in our C compiler. Contrary to the statement 
on page 16 of the "C Reference Manusi”, "...The length of the array is 
taken to be the maximum of the number of expressions in the list and 

the square-bracketed constant in the array's declarator. -+.", we found 
that, in the case of an array initialized to a character string, the 
length of the array vas set to the length of the character string, rather 
then the greater value in the square brackets. We have not determined 
either the cause of the error or if the same thing occurs on other types 
of initialization. 


I hope this information is of some use. 


Ms 
err 





‘DMP / du 


ae 


Fee Geer ty of Tawa ar Austin « Teas Student Pupicetions Building 3.200 





September 147 1976 
Subject! Floatins point simulator in I/D space 


We have found that the flostins point simulator packase does 
not orerate promerly with I/D space seraration., The difficultw is 
caused (1) by fetchina the flostins roint instruction from the 
program, and (2) incorrectlw interpreting such addressing modes 
as immediate and absolute, The chanses to fix this reeuire being 
able to access the I sesce as data. It is our suess thet the MFPI 
instruction can be used for this surrose. Even sor rather extensive 
modifications tothe flostins roint interpreter are reauired. We 
would like toknow if anwone else has discovered and fined this eroblem, 


Alsos we find that cdb will not operate rrorerly if floating roint 
simulation is beins used, end have corrected the problem. We would 
be #liad to let anyone know about this who cares. 


Alar G. Nemeth 

M.I.T. Lincoln Laboratory 
_ Lexinstony Ma. 02173 
(617)862-5500 «696 or x212