The length of the document below is: 4 page(s) long
The self-declared author(s) is/are:
www.cs.dartmouth.edu
The subject is as follows:
Subject: Original authors did not specify.
The original URL is: LINK
The access date was:
Access date: 2019-04-12 14:35:23.387795
Please be aware that this may be under copyright restrictions. Please send an email to admin@pharmacoengineering.com for any AI-generated issues.
Loading...
The content is as follows:
Communication Files:
Interprocess IO bef
ore Pipes
M. Douglas McIlr
oyDartmouth Colle
gedoug@cs.dartmouth.eduFebruary, 2017
IntroductionSome time after the introduction of pipes to Unix, we in the Bell Labs Unix lab learned
that the Dartmouth T
ime-Sharing System (DTSS) had a mechanism for process-to-
process IO called
communication„les
. Unfortunately we didn’
t kno
w e
xactly ho
w the
yworked. When I retired from Bell Labs to Dartmouth in 1997, I ask
ed around fruitlessly
for further information. At last, at a DTSS reunion or
ganized by T
om K
urtz, who had fos-
tered the project, I met Sidne
y Marshall, who had been in
volved in the implementation.
He e
xplained the concept. The picture w
as rounded out in discussion with another partici-
pant, Stephen Garland, who had edited the
DTSSPr
ogramming Manual
.Communication „les were much more complicated than Unix pipes. The
y were also
more po
werful. Pipes could be simulated by communication „les, b
ut not vice v
ersa. A
pipe can handle neither the tw
o-way communication nor the out-of-band signaling that
communication „les support.
The programming manual’
s description of communication „les ran to man
y pages. [a
vail-able at http://www
munication „les remained be
yond the w
orking toolkit e
ven of man
y DTSS insiders. Ne
v-ertheless communication „les played an indispensable role: one or more communication
„les mediated e
very user’
s interaction with the system.
Much of the detail belo
w comes from collections of DTSS documents that Garland and
Marshall ha
ve deposited with the Dartmouth library
. Further information w
as gleaned
from an email con
versation among DTSS alumni, to which Peter Do
yle kindly introduced
me.DatesCommunication „les signi„cantly antedated Unix pipes. Evidence from design docu-
ments puts the origin of the concept sometime in 1967, between the writing of an outline
of features for the Phase II DTSS system dated March, which doesn’
t mention communi-
cation „les, and a summary of e
xecutive services dated 29 August, which does. As com-
munication „les were used for terminal sessions, the
y were operational when the system
went li
ve on January 6, 1969 [John K
emeny, January 20, 1969]. (More than three years
before pipes deb
Please note all content on this page was automatically generated via our AI-based algorithm (BishopKingdom ID: 10Tq1v4OxHDDUoz1cQO5). Please let us know if you find any errors.
Reload document 
Open in new tab