2000
%!PS-Adobe-2.0
%%Copyright: Copyright (c) 1993 AT&T, All Rights Reserved
%%Version: 3.4
%%DocumentFonts: (atend)
%%Pages: (atend)
%%BoundingBox: (atend)
%%EndComments
/DpostDict 200 dict def
DpostDict begin
%
% Copyright (c) 1993 AT&T, All Rights Reserved
%
% Version 3.4 prologue for troff files.
%

/#copies 1 store
/Prologue (dpost.ps) def
/aspectratio 1 def
/formsperpage 1 def
/landscape false def
/linewidth .3 def
/magnification 1 def
/margin 0 def
/orientation 0 def
/resolution 720 def
/rotation 1 def
/xoffset 0 def
/yoffset 0 def

/roundpage true def
/useclippath true def
/pagebbox [0 0 612 792] def

/R  /Times-Roman def
/I  /Times-Italic def
/B  /Times-Bold def
/BI /Times-BoldItalic def
/H  /Helvetica def
/HI /Helvetica-Oblique def
/HB /Helvetica-Bold def
/HX /Helvetica-BoldOblique def
/CW /Courier def
/CO /Courier def
/CI /Courier-Oblique def
/CB /Courier-Bold def
/CX /Courier-BoldOblique def
/PA /Palatino-Roman def
/PI /Palatino-Italic def
/PB /Palatino-Bold def
/PX /Palatino-BoldItalic def
/Hr /Helvetica-Narrow def
/Hi /Helvetica-Narrow-Oblique def
/Hb /Helvetica-Narrow-Bold def
/Hx /Helvetica-Narrow-BoldOblique def
/KR /Bookman-Light def
/KI /Bookman-LightItalic def
/KB /Bookman-Demi def
/KX /Bookman-DemiItalic def
/AR /AvantGarde-Book def
/AI /AvantGarde-BookOblique def
/AB /AvantGarde-Demi def
/AX /AvantGarde-DemiOblique def
/NR /NewCenturySchlbk-Roman def
/NI /NewCenturySchlbk-Italic def
/NB /NewCenturySchlbk-Bold def
/NX /NewCenturySchlbk-BoldItalic def
/ZD /ZapfDingbats def
/ZI /ZapfChancery-MediumItalic def
/S  /S def
/S1 /S1 def
/GR /Symbol def

/inch {72 mul} bind def
/min {2 copy gt {exch} if pop} bind def

/setup {
	counttomark 2 idiv {def} repeat pop

	landscape {/orientation 90 orientation add def} if
	/scaling 72 resolution div def
	linewidth setlinewidth
	1 setlinecap

	pagedimensions
	xcenter ycenter translate
	orientation rotation mul rotate
	width 2 div neg height 2 div translate
	xoffset inch yoffset inch neg translate
	margin 2 div dup neg translate
	magnification dup aspectratio mul scale
	scaling scaling scale

	addmetrics
	0 0 moveto
} def

/pagedimensions {
	useclippath userdict /gotpagebbox known not and {
		/pagebbox [clippath pathbbox newpath] def
		roundpage currentdict /roundpagebbox known and {roundpagebbox} if
	} if
	pagebbox aload pop
	4 -1 roll exch 4 1 roll 4 copy
	landscape {4 2 roll} if
	sub /width exch def
	sub /height exch def
	add 2 div /xcenter exch def
	add 2 div /ycenter exch def
	userdict /gotpagebbox true put
} def

/landscapepage {
	landscape not {
		0 height scaling div neg translate	% not quite
		90 rotate
	} if
} bind def

/portraitpage {
	landscape {
		width scaling div 0 translate	% not quite
		-90 rotate
	} if
} bind def

/addmetrics {
	/Symbol /S null Sdefs cf
	/Times-Roman /S1 StandardEncoding dup length array copy S1defs cf
} def

/pagesetup {
	/page exch def
	currentdict /pagedict known currentdict page known and {
		page load pagedict exch get cvx exec
	} if
} def

/decodingdefs [
	{counttomark 2 idiv {y moveto show} repeat}
	{neg /y exch def counttomark 2 idiv {y moveto show} repeat}
	{neg moveto {2 index stringwidth pop sub exch div 0 32 4 -1 roll widthshow} repeat}
	{neg moveto {spacewidth sub 0.0 32 4 -1 roll widthshow} repeat}
	{counttomark 2 idiv {y moveto show} repeat}
	{neg setfunnytext}
] def

/setdecoding {/t decodingdefs 3 -1 roll get bind def} bind def

/w {neg moveto show} bind def
/m {neg dup /y exch def moveto} bind def
/done {/lastpage where {pop lastpage} if} def

/f {
	dup /font exch def findfont exch
	dup /ptsize exch def scaling div dup /size exch def scalefont setfont
	linewidth ptsize mul scaling 10 mul div setlinewidth
	/spacewidth ( ) stringwidth pop def
} bind def

/changefont {
	/fontheight exch def
	/fontslant exch def
	currentfont [
		1 0
		fontheight ptsize div fontslant sin mul fontslant cos div
		fontheight ptsize div
		0 0
	] makefont setfont
} bind def

/sf {f} bind def

/cf {
	dup length 2 idiv
	/entries exch def
	/chtab exch def
	/newencoding exch def
	/newfont exch def

	findfont dup length 1 add dict
	/newdict exch def
	{1 index /FID ne {newdict 3 1 roll put}{pop pop} ifelse} forall

	newencoding type /arraytype eq {newdict /Encoding newencoding put} if

	newdict /Metrics entries dict put
	newdict /Metrics get
	begin
		chtab aload pop
		1 1 entries {pop def} for
		newfont newdict definefont pop
	end
} bind def

%
% A few arrays used to adjust reference points and character widths in some
% of the printer resident fonts. If square roots are too high try changing
% the lines describing /radical and /radicalex to,
%
%	/radical	[0 -75 550 0]
%	/radicalex	[-50 -75 500 0]
%
% Move braceleftbt a bit - default PostScript character is off a bit.
%

/Sdefs [
	/bracketlefttp		[201 500]
	/bracketleftbt		[201 500]
	/bracketrighttp		[-81 380]
	/bracketrightbt		[-83 380]
	/braceleftbt		[203 490]
	/bracketrightex		[220 -125 500 0]
	/radical		[0 0 550 0]
	/radicalex		[-50 0 500 0]
	/parenleftex		[-20 -170 0 0]
	/integral		[100 -50 500 0]
	/infinity		[10 -75 730 0]
] def

/S1defs [
	/underscore		[0 80 500 0]
	/endash			[7 90 650 0]
] def
end
%%EndProlog
%%BeginSetup
DpostDict begin
mark
/rotation 1 def
/gotpagebbox true def
/linewidth 0.5 def
/xoffset 0 def
/yoffset 0 def
/#copies 1 store
/magnification 1 def
%%FormsPerPage: 1
/formsperpage 1 def
%%Patch from lp
%%EndPatch from lp
/landscape false def
/resolution 720 def
setup
2 setdecoding
end
%%EndSetup
%%Page: 1 1
%%PageBoundingBox: (atend)
DpostDict begin
/saveobj save def
mark
1 pagesetup
12 PB f
(The Evolution of the Unix Time-sharing System*)6 2680 1 1540 1230 t
10 PI f
(Dennis M. Ritchie)2 751 1 2504 1470 t
10 PA f
(Bell Laboratories)1 750 1 2505 1650 t
(Murray Hill, New Jersey 07974)4 1367 1 2196 1770 t
10 PI f
(ABSTRACT)2621 2150 w
10 PA f
( early development of the Unix)5 1456(This paper presents a brief history of the)7 1894 2 1330 2410 t
( the file system, the)4 995( concentrates on the evolution of)5 1626( It)1 154(operating system.)1 825 4 1080 2530 t
( atten-)1 294( Some)1 308(process-control mechanism, and the idea of pipelined commands.)7 2998 3 1080 2650 t
(tion is paid to social conditions during the development of the system.)11 3110 1 1080 2770 t
10 PB f
(Introduction)720 3166 w
10 PA f
( Unix operating system has come into wide use, so wide that)11 2722(During the past few years, the)5 1348 2 970 3322 t
( important characteristics have)3 1401( Its)1 173( trademark of Bell Laboratories.)4 1460(its very name has become a)5 1286 4 720 3442 t
( much rewriting and tinkering since the first pub-)8 2211( has suffered)2 569( It)1 119(become known to many people.)4 1421 4 720 3562 t
( Unix was born in 1969)5 1027( However,)1 492(lication describing it in 1974 [1], but few fundamental changes.)9 2801 3 720 3682 t
( story.)1 280(not 1974, and the account of its development makes a little-known and perhaps instructive)13 4040 2 720 3802 t
(This paper presents a technical and social history of the evolution of the system.)13 3529 1 720 3922 t
10 PB f
(Origins)720 4162 w
10 PA f
( 1968-1969 was somewhat unsettled.)4 1654(For computer science at Bell Laboratories, the period)7 2416 2 970 4318 t
( the)1 164(The main reason for this was the slow, though clearly inevitable, withdrawal of the Labs from)15 4156 2 720 4438 t
( as a whole, the problem was the increasing)8 1986( the Labs computing community)4 1469( To)1 173(Multics project.)1 692 4 720 4558 t
( let alone the)3 579(obviousness of the failure of Multics to deliver promptly any sort of usable system,)13 3741 2 720 4678 t
( of this time, the Murray Hill Computer Center was also)10 2597( much)1 286( For)1 212(panacea envisioned earlier.)2 1225 4 720 4798 t
( shake-up that)2 633( Another)1 421( simulated the GE 635.)4 995(running a costly GE 645 machine that inadequately)7 2271 4 720 4918 t
( organizational separation of computing services and com-)7 2675(occurred d
2000
uring this period was the)5 1645 2 720 5038 t
(puting research.)1 716 1 720 5158 t
( point of view of the group that was to be most involved in the beginnings of Unix)17 3666(From the)1 404 2 970 5314 t
( the decline and fall of Multics had a directly)9 1970(\(K. Thompson, Ritchie, M. D. McIlroy, J. F. Ossanna\),)8 2350 2 720 5434 t
( so we)2 280( were among the last Bell Laboratories holdouts actually working on Multics,)11 3410( We)1 198(felt effect.)1 432 4 720 5554 t
( interactive computing)2 1013( important, the convenient)3 1193( More)1 295(still felt some sort of stake in its success.)8 1819 4 720 5674 t
( to our limited)3 671(service that Multics had promised to the entire community was in fact available)12 3649 2 720 5794 t
( used to develop Multics, and later under Multics itself.)9 2571(group, at first under the CTSS system)6 1749 2 720 5914 t
( could support us, albeit at exorbitant)6 1680(Even though Multics could not then support many users, it)9 2640 2 720 6034 t
( the pleasant niche we occupied, because no similar ones were avail-)11 3047( didn't want to lose)4 872(cost. We)1 401 3 720 6154 t
( the time-sharing service that would later be offered under GE's operating system did)13 3872(able; even)1 448 2 720 6274 t
( just a good environment in which to do program-)9 2216( we wanted to preserve was not)6 1411( What)1 292(not exist.)1 401 4 720 6394 t
( knew from experience that the)5 1429( We)1 209( could form.)2 562(ming, but a system around which a fellowship)7 2120 4 720 6514 t
( time-shared machines, is not just)5 1487(essence of communal computing, as supplied by remote-access,)7 2833 2 720 6634 t
8 S1 f
(__________________)720 6734 w
8 PA f
( Programming Methodology conference at Sydney,)5 1824(*This paper was first presented at the Language Design and)9 2136 2 720 6834 t
( conference proceedings were published as)5 1516( The)1 174(Australia, September 1979.)2 946 3 720 6934 t
8 PI f
(Lecture Notes in Computer Science #79:)5 1303 1 3377 6934 t
(Language Design and Programming Methodology,)4 1682 1 720 7034 t
8 PA f
( rendition is based on a reprinted)6 1243( This)1 204(Springer-Verlag, 1980.)1 800 3 2433 7034 t
(version appearing in AT&T Bell Laboratories Technical Journal)7 2226 1 720 7134 t
8 PB f
(63)2966 7134 w
8 PA f
(No. 6 Part 2, October 1984, pp. 1577-93.)7 1380 1 3066 7134 t
cleartomark
showpage
saveobj restore
end
%%PageBoundingBox: 61 64 514 691
%%EndPage: 1 1
%%Page: 2 2
%%PageBoundingBox: (atend)
DpostDict begin
/saveobj save def
mark
2 pagesetup
10 PA f
(- 2 -)2 166 1 2797 480 t
(to type programs into a terminal instead of a keypunch, but to encourage close communication.)14 4219 1 720 840 t
( search took sev-)3 756( The)1 224( began trying to find an alternative to Multics.)8 2091(Thus, during 1969, we)3 999 4 970 996 t
( we \(mainly Ossanna, Thompson, Ritchie\) lobbied intensively for the)9 3031( 1969)1 226( Throughout)1 587(eral forms.)1 476 4 720 1116 t
( for which we promised to write an operating system; the)10 2645(purchase of a medium-scale machine)4 1675 2 720 1236 t
( effort was)2 473( The)1 220( were the DEC PDP-10 and the SDS \(later Xerox\) Sigma 7.)11 2575(machines we suggested)2 1052 4 720 1356 t
( were never clearly and finally turned down, but yet were cer-)11 2799(frustrating, because our proposals)3 1521 2 720 1476 t
( final blow to this)4 783( The)1 221( times it seemed we were very near success.)8 1962( Several)1 378(tainly never accepted.)2 976 5 720 1596 t
( minimize finan-)2 734(effort came when we presented an exquisitely complicated proposal, designed to)10 3586 2 720 1716 t
( some third-party lease, and a plan to turn in a)10 2090(cial outlay, that involved some outright purchase,)6 2230 2 720 1836 t
( proposal was)2 634( The)1 225(DEC KA-10 processor on the soon-to-be-announced and more capable KI-10.)9 3461 3 720 1956 t
( \(then vice-president of Research\) had reacted to)7 2147(rejected, and rumor soon had it that W. O. Baker)9 2173 2 720 2076 t
(it with the comment `Bell Laboratories just doesn't do business this way!')11 3228 1 720 2196 t
( the time\) that we)4 831(Actually, it is perfectly obvious in retrospect \(and should have been at)11 3239 2 970 2352 t
( More-)1 324(were asking the Labs to spend too much money on too few people with too vague a plan.)17 3996 2 720 2472 t
( systems were not, for our management, an attrac-)8 2225(over, I am quite sure that at that time operating)9 2095 2 720 2592 t
( process of extricating themselves not only)6 1907( were in the)3 538( They)1 279(tive area in which to support work.)6 1596 4 720 2712 t
(from an operating system development effort that had failed, but from running the local Compu-)14 4320 1 720 2832 t
( have seemed that buying a machine such as we suggested might lead)12 3121( it may)2 314( Thus)1 275(tation Center.)1 610 4 720 2952 t
( one hand to yet another Multics, or on the other, if we produced something useful, to yet)17 4039(on the)1 281 2 720 3072 t
(another Comp Center for them to be responsible for.)8 2319 1 720 3192 t
( work also.)2 535(Besides the financial agitations that took place in 1969, there was technical)11 3535 2 970 3348 t
( notes, the basic)3 707(Thompson, R. H. Canaday, and Ritchie developed, on blackboards and scribbled)10 3613 2 720 3468 t
( of the design was)4 880( Most)1 296( to become the heart of Unix.)6 1400(design of a file system that was later)7 1744 4 720 3588 t
( but I believe I contributed the)6 1364(Thompson's, as was the impulse to think about file systems at all,)11 2956 2 720 3708 t
( itch for creation of an operating system took several forms dur-)11 2866( Thompson's)1 601( files.)1 235(idea of device)2 618 4 720 3828 t
( the performance of the)4 1054(ing this period; he also wrote \(on Multics\) a fairly detailed simulation of)12 3266 2 720 3948 t
( started work on)3 732( addition, he)2 557( In)1 142(proposed file system design and of paging behavior of programs.)9 2889 4 720 4068 t
( GE-645, going as far as writing an assembler for the machine and)12 2931(a new operating system for the)5 1389 2 720 4188 t
( was to)2 319(a rudimentary operating system kernel whose greatest achievement, so far as I remember,)12 4001 2 720 4308 t
( mere message was)3 899( complexity of the machine was such that a)8 2043( The)1 235(type a greeting message.)3 1143 4 720 4428 t
( accomplishment, but when it became clear that the lifetime of the 645 at)13 3278(already a fairly notable)3 1042 2 720 4548 t
(the Labs was measured in months, the work was dropped.)9 2588 1 720 4668 t
( written on Mul-)3 753( First)1 258( developed the game of `Space Travel.')6 1753(Also during 1969, Thompson)3 1306 4 970 4824 t
( GECOS \(the operating system for the GE, later Honey-)9 2536(tics, then transliterated into Fortran for)5 1784 2 720 4944 t
( it was nothing less than a simulation of the movement of the major bodies of the Solar)17 3871(well, 635\),)1 449 2 720 5064 t
( attempting to)2 636(System, with the player guiding a ship here and there, observing the scenery, and)13 3684 2 720 5184 t
( GECOS version was unsatisfactory in two important)7 2354( The)1 218( various planets and moons.)4 1248(land on the)2 500 4 720 5304 t
( was jerky and hard to control because one had)9 2107(respects: first, the display of the state of the game)9 2213 2 720 5424 t
( It)1 121( the big computer.)3 819(to type commands at it, and second, a game cost about $75 for CPU time on)15 3380 3 720 5544 t
( therefore, for Thompson to find a little-used PDP-7 computer with an excellent)12 3529(did not take long,)3 791 2 720 5664 t
( and I rewrote Space)4 933( He)1 188( as a Graphic-II terminal.)4 1128(display processor; the whole system was used)6 2071 4 720 5784 t
( because)1 373( undertaking was more ambitious than it might seem;)8 2388( The)1 219(Travel to run on this machine.)5 1340 4 720 5904 t
( a floating-point arithmetic package, the point-)6 2065(we disdained all existing software, we had to write)8 2255 2 720 6024 t
( that con-)2 420(wise specification of the graphic characters for the display, and a debugging subsystem)12 3900 2 720 6144 t
(
2000
 this was writ-)3 636( All)1 188( screen.)1 332(tinuously displayed the contents of typed-in locations in a corner of the)11 3164 4 720 6264 t
( cross-assembler that ran under GECOS and produced paper tapes)9 2959(ten in assembly language for a)5 1361 2 720 6384 t
(to be carried to the PDP-7.)5 1160 1 720 6504 t
( served mainly as an introduction to)6 1664(Space Travel, though it made a very attractive game,)8 2406 2 970 6660 t
( began implement-)2 835( Thompson)1 505( Soon)1 274(the clumsy technology of preparing programs for the PDP-7.)8 2706 4 720 6780 t
( file system' would be more accurate\) that had been)9 2430(ing the paper file system \(perhaps `chalk)6 1890 2 720 6900 t
( he pro-)2 368( file system without a way to exercise it is a sterile proposition, so)13 3061( A)1 141(designed earlier.)1 750 4 720 7020 t
( particular)1 469(ceeded to flesh it out with the other requirements for a working operating system, in)14 3851 2 720 7140 t
( came a small set of user-level utilities: the means to copy, print,)12 2961( Then)1 287(the notion of processes.)3 1072 3 720 7260 t
cleartomark
showpage
saveobj restore
end
%%PageBoundingBox: 61 50 514 764
%%EndPage: 2 2
%%Page: 3 3
%%PageBoundingBox: (atend)
DpostDict begin
/saveobj save def
mark
3 pagesetup
10 PA f
(- 3 -)2 166 1 2797 480 t
( this time all the)4 731( to)1 118( Up)1 193(delete, and edit files, and of course a simple command interpreter \(shell\).)11 3278 4 720 840 t
( transferred to the PDP-7 on paper tape; but)8 1975(programs were written using GECOS and files were)7 2345 2 720 960 t
( it was not)3 506( Although)1 493( to support itself.)3 800(once an assembler was completed the system was able)8 2521 4 720 1080 t
( suggested the name `Unix,' in a somewhat treacherous)8 2479(until well into 1970 that Brian Kernighan)6 1841 2 720 1200 t
(pun on `Multics,' the operating system we know today was born.)10 2877 1 720 1320 t
10 PB f
(The PDP-7 Unix file system)4 1259 1 720 1560 t
10 PA f
( had)1 194( It)1 117(Structurally, the file system of PDP-7 Unix was nearly identical to today's.)11 3274 3 970 1716 t
( i-list: a linear array of)5 1038(1\) An)1 386 2 720 1872 t
10 PI f
(i-nodes)2182 1872 w
10 PA f
( i-node contained less than it)5 1337( An)1 200(each describing a file.)3 993 3 2510 1872 t
( but the essential information was the same: the protection mode of the file, its)14 3606(does now,)1 464 2 970 1992 t
(type and size, and the list of physical blocks holding the contents.)11 2894 1 970 2112 t
( of file containing a sequence of names and the associated i-)11 2866( a special kind)3 694(2\) Directories:)1 760 3 720 2268 t
(number.)970 2388 w
( the)1 168( device specification was not contained explicitly in)7 2289( The)1 220( files describing devices.)3 1080(3\) Special)1 563 5 720 2544 t
( specific i-numbers corresponded to specific)5 1946(i-node, but was instead encoded in the number:)7 2124 2 970 2664 t
(files.)970 2784 w
( write, open, creat)3 830( Read,)1 313(The important file system calls were also present from the start.)10 2927 3 970 2940 t
( with one very important exception, discussed below, they were similar to what one)13 3835(\(sic\), close:)1 485 2 720 3060 t
( the word, not the byte, because the)7 1615( minor difference was that the unit of I/O was)9 2096( A)1 134(finds now.)1 475 4 720 3180 t
( practice this meant merely that all programs dealing)8 2404( In)1 150(PDP-7 was a word-addressed machine.)4 1766 3 720 3300 t
( streams ignored null characters, because null was used to pad a file to an even)15 3673(with character)1 647 2 720 3420 t
( minor, occasionally annoying difference was the lack of erase and)10 2936( Another)1 421(number of characters.)2 963 3 720 3540 t
( few pro-)2 430( a)1 88( Only)1 285( in effect, were always in raw mode.)7 1683( Terminals,)1 533(kill processing for terminals.)3 1301 6 720 3660 t
(grams \(notably the shell and the editor\) bothered to implement erase-kill processing.)11 3742 1 720 3780 t
( the PDP-7 file system was in)6 1297(In spite of its considerable similarity to the current file system,)10 2773 2 970 3936 t
( different: there were no path names, and each file-name argument to the)12 3380(one way remarkably)2 940 2 720 4056 t
( the)1 182( in)1 129( Links,)1 338(system was a simple name \(without `/'\) taken relative to the current directory.)12 3671 4 720 4176 t
( set of conventions, they were the princi-)7 1828( with an elaborate)3 812( Together)1 455(usual Unix sense, did exist.)4 1225 4 720 4296 t
(pal means by which the lack of path names became acceptable.)10 2770 1 720 4416 t
(The)970 4572 w
10 PI f
(link)1162 4572 w
10 PA f
(call took the form)3 781 1 1343 4572 t
(link\(dir, file, newname\))2 1040 1 1080 4752 t
(where)720 4932 w
10 PI f
(dir)1026 4932 w
10 PA f
( directory,)1 461(was a directory file in the current)6 1493 2 1172 4932 t
10 PI f
(file)3156 4932 w
10 PA f
(an existing entry in that directory, and)6 1731 1 3309 4932 t
10 PI f
(newname)720 5052 w
10 PA f
( Because)1 410(the name of the link, which was added to the current directory.)11 2868 2 1137 5052 t
10 PI f
(dir)4447 5052 w
10 PA f
(needed to)1 444 1 4596 5052 t
( directory, it is evident that today's prohibition against links to directories was)12 3569(be in the current)3 751 2 720 5172 t
(not enforced; the PDP-7 Unix file system had the shape of a general directed graph.)14 3678 1 720 5292 t
( link to all directories of interest, there existed)8 2053(So that every user did not need to maintain a)9 2017 2 970 5448 t
(a directory called)2 775 1 720 5568 t
10 PI f
(dd)1524 5568 w
10 PA f
( to make a link to)5 780( Thus,)1 299(that contained entries for the directory of each user.)8 2308 3 1653 5568 t
(file)720 5688 w
10 PI f
(x)884 5688 w
10 PA f
(in directory)1 518 1 959 5688 t
10 PI f
(ken)1502 5688 w
10 PA f
(, I might do)3 514 1 1641 5688 t
(ln dd ken ken)3 608 1 1080 5868 t
(ln ken x x)3 428 1 1080 5988 t
(rm ken)1 315 1 1080 6108 t
( practice.)1 402(This scheme rendered subdirectories sufficiently hard to use as to make them unused in)13 3918 2 720 6288 t
( to create a directory while the system was)8 1909(Another important barrier was that there was no way)8 2411 2 720 6408 t
( made during recreation of the file system from paper tape, so that directories)13 3543(running; all were)2 777 2 720 6528 t
(were in effect a nonrenewable resource.)5 1752 1 720 6648 t
(The)970 6804 w
10 PI f
(dd)1173 6804 w
10 PA f
(convention made the)2 952 1 1309 6804 t
10 PI f
(chdir)2297 6804 w
10 PA f
( took multiple argu-)3 925( It)1 128(command relatively convenient.)2 1446 3 2541 6804 t
( Thus)1 271(ments, and switched the current directory to each named directory in turn.)11 3306 2 720 6924 t
(chdir dd ken)2 566 1 1080 7104 t
(would move to directory)3 1131 1 720 7284 t
10 PI f
(ken)1885 7284 w
10 PA f
(. \(Incidentally,)1 671 1 2024 7284 t
10 PI f
(chdir)2729 7284 w
10 PA f
(was spelled)1 526 1 2971 7284 t
10 PI f
(ch)3531 7284 w
10 PA f
( this was expanded when)4 1162(; why)1 256 2 3622 7284 t
cleartomark
showpage
saveobj restore
end
%%PageBoundingBox: 61 48 514 764
%%EndPage: 3 3
%%Page: 4 4
%%PageBoundingBox: (atend)
DpostDict begin
/saveobj save def
mark
4 pagesetup
10 PA f
(- 4 -)2 166 1 2797 480 t
(we went to the PDP-11 I don't remember.\))7 1867 1 720 840 t
( inconvenience of the implementation of the file system, aside from the)11 3291(The most serious)2 779 2 970 996 t
( mentioned, directories and)3 1216(lack of path names, was the difficulty of changing its configuration; as)11 3104 2 720 1116 t
( of a new device was)5 938( Installation)1 555( disk was recreated.)3 890(special files were both made only when the)7 1937 4 720 1236 t
( the system; for example)4 1076(very painful, because the code for devices was spread widely throughout)10 3244 2 720 1356 t
( surprisingly, there was no notion of)6 1613( Not)1 221(there were several loops that visited each device in turn.)9 2486 3 720 1476 t
(mounting a removable disk pack, because the machine had only a single fixed-head disk.)13 3930 1 720 1596 t
( a drasti
2000
cally simplified)3 1075(The operating system code that implemented this file system was)9 2995 2 970 1752 t
( important simplification followed from the fact that the sys-)9 2718( One)1 241(version of the present scheme.)4 1361 3 720 1872 t
( in memory at a time, and control was)8 1772(tem was not multi-programmed; only one program was)7 2548 2 720 1992 t
( for example, there was an)5 1188( So,)1 189(passed between processes only when an explicit swap took place.)9 2943 3 720 2112 t
10 PI f
(iget)720 2232 w
10 PA f
( i-node available, but it left the i-node in a constant, static location)12 2936(routine that made a named)4 1207 2 897 2232 t
( precursor of the current)4 1131( A)1 141( a large table of active i-nodes.)6 1415(rather than returning a pointer into)5 1633 4 720 2352 t
( of)1 122(buffering mechanism was present \(with about 4 buffers\) but there was essentially no overlap)13 4198 2 720 2472 t
( disk attached to the)4 892( The)1 217( simplicity.)1 489( was avoided not merely for)5 1248( This)1 241(disk I/O with computation.)3 1233 6 720 2592 t
( the other)2 447( On)1 199( word every 2 microseconds.)4 1316(PDP-7 was fast for its time; it transferred one 18-bit)9 2358 4 720 2712 t
( of 1 microsecond, and most instructions took 2)8 2150(hand, the PDP-7 itself had a memory cycle time)8 2170 2 720 2832 t
( indirectly addressed instruc-)3 1297( However,)1 489( the instruction itself, one for the operand\).)7 1899(cycles \(one for)2 635 4 720 2952 t
( common, because the machine had no index)7 2063(tions required 3 cycles, and indirection was quite)7 2257 2 720 3072 t
( The)1 225( access memory during an instruction.)5 1728( the DMA controller was unable to)6 1575(registers. Finally,)1 792 4 720 3192 t
( incur overrun errors if any indirectly-addressed instructions)7 2823(upshot was that the disk would)5 1497 2 720 3312 t
( control could not be returned to the user, nor in)10 2222( Thus)1 280( while it was transferring.)4 1172(were executed)1 646 4 720 3432 t
( the)1 165( interrupt routines for)3 975( The)1 219(fact could general system code be executed, with the disk running.)10 2961 4 720 3552 t
( very strange)2 599(clock and terminals, which needed to be runnable at all times, had to be coded in)15 3721 2 720 3672 t
(fashion to avoid indirection.)3 1248 1 720 3792 t
10 PB f
(Process control)1 685 1 720 4032 t
10 PA f
( by which processes are created and used;)7 1993(By `process control,' I mean the mechanisms)6 2077 2 970 4188 t
(today the system calls)3 993 1 720 4308 t
10 PI f
(fork)1745 4308 w
10 PA f
(,)1900 4308 w
10 PI f
(exec)1957 4308 w
10 PA f
(,)2126 4308 w
10 PI f
(wait)2183 4308 w
10 PA f
(, and)1 226 1 2360 4308 t
10 PI f
(exit)2617 4308 w
10 PA f
( the file sys-)3 544( Unlike)1 354(implement these mechanisms.)2 1344 3 2798 4308 t
( form from the earliest days, the process control scheme)9 2542(tem, which existed in nearly its present)6 1778 2 720 4428 t
( introduction of path)3 919( \(The)1 250(underwent considerable mutation after PDP-7 Unix was already in use.)9 3151 3 720 4548 t
( advance, but not a change in)6 1305(names in the PDP-11 system was certainly a considerable notational)9 3015 2 720 4668 t
(fundamental structure.\))1 1051 1 720 4788 t
( by the shell can be summarized as fol-)8 1805(Today, the way in which commands are executed)7 2265 2 970 4944 t
(lows:)720 5064 w
( shell reads a command line from the terminal.)8 2060(1\) The)1 417 2 720 5220 t
( creates a child process by)5 1143(2\) It)1 317 2 720 5376 t
10 PI f
(fork.)2205 5376 w
10 PA f
( child process uses)3 819(3\) The)1 417 2 720 5532 t
10 PI f
(exec)1981 5532 w
10 PA f
(to call in the command from a file.)7 1515 1 2175 5532 t
( uses)1 222( the parent shell)3 721(4\) Meanwhile,)1 773 3 720 5688 t
10 PI f
(wait)2466 5688 w
10 PA f
(to wait for the child \(command\) process to terminate)8 2367 1 2673 5688 t
(by calling)1 431 1 970 5808 t
10 PI f
(exit.)1426 5808 w
10 PA f
( parent shell goes back to step 1\).)7 1455(5\) The)1 417 2 720 5964 t
( were)1 251( There)1 312( Unix.)1 274(Processes \(independently executing entities\) existed very early in PDP-7)8 3233 4 970 6120 t
( There)1 314(in fact precisely two of them, one for each of the two terminals attached to the machine.)16 4006 2 720 6240 t
(was no)1 316 1 720 6360 t
10 PI f
(fork)1064 6360 w
10 PA f
(,)1219 6360 w
10 PI f
(wait)1272 6360 w
10 PA f
(, or)1 148 1 1449 6360 t
10 PI f
(exec)1625 6360 w
10 PA f
( was an)2 341(. There)1 334 2 1794 6360 t
10 PI f
(exit)2498 6360 w
10 PA f
(, but its meaning was rather different, as will be seen.)10 2392 1 2648 6360 t
(The main loop of the shell went as follows.)8 1889 1 720 6480 t
( then opened the terminal special file for standard input)9 2569( shell closed all its open files,)6 1334(1\) The)1 417 3 720 6636 t
(and output \(file descriptors 0 and 1\).)6 1613 1 970 6756 t
( read a command line from the terminal.)7 1787(2\) It)1 317 2 720 6912 t
( it)1 90( Then)1 278( to the file specifying the command, opened the file, and removed the link.)13 3327( linked)1 308(3\) It)1 317 5 720 7068 t
( to it; this bootstrap)4 908(copied a small bootstrap program to the top of memory and jumped)11 3162 2 970 7188 t
( location of the)3 736(program read in the file over the shell code, then jumped to the first)13 3334 2 970 7308 t
cleartomark
showpage
saveobj restore
end
%%PageBoundingBox: 61 45 514 764
%%EndPage: 4 4
%%Page: 5 5
%%PageBoundingBox: (atend)
DpostDict begin
/saveobj save def
mark
5 pagesetup
10 PA f
(- 5 -)2 166 1 2797 480 t
(command \(in effect an)3 986 1 970 840 t
10 PI f
(exec)1981 840 w
10 PA f
(\).)2150 840 w
( work, then terminated by calling)5 1495( command did its)3 783(4\) The)1 417 3 720 996 t
10 PI f
(exit)3444 996 w
10 PA f
(. The)1 246 1 3594 996 t
10 PI f
(exit)3869 996 w
10 PA f
(call caused the system)3 992 1 4048 996 t
(to read in a fresh copy of the shell over the terminated command, then to jump to its start)18 4070 1 970 1116 t
(\(and thus in effect to go to step 1\).)8 1499 1 970 1236 t
( degree to which it)4 883(The most interesting thing about this primitive implementation is the)9 3187 2 970 1392 t
( it could support neither background pro-)6 1908( True,)1 294( more fully later.)3 768(anticipated themes developed)2 1350 4 720 1512 t
( alone pipes and filters\); but IO redirection \(via `)9 2145(cesses nor shell command files \(let)5 1523 2 720 1632 t
10 S f
(<)4388 1632 w
10 PA f
(' and `)2 279 1 4443 1632 t
10 S f
(>)4722 1632 w
10 PA f
('\) was)1 263 1 4777 1632 t
( implementation of redirection was quite straightforward;)6 2606( The)1 225(soon there; it is discussed below.)5 1489 3 720 1752 t
( Cru-)1 254(in step 3\) above the shell just replaced its standard input or output with the appropriate file.)16 4066 2 720 1872 t
( implementation of the shell as a user-level program)8 2455(cial to subsequent development was the)5 1865 2 720 1992 t
(stored in a file, rather than a part of the operating system.)11 2546 1 720 2112 t
(The structure of this process control scheme, with one process per terminal, is similar to)14 4070 1 970 2268 t
( many interactive systems, for example CTSS, Multics, Honeywell TSS, and IBM TSS and)13 4024(that of)1 296 2 720 2388 t
( general such systems require special mechanisms to implement useful facilities such as)12 3953(TSO. In)1 367 2 720 2508 t
( and command files; Unix at that stage didn't bother to supply the special)13 3281(detached computations)1 1039 2 720 2628 t
( example, a newly)3 862( For)1 222( problems.)1 488( also exhibited some irritating, idiosyncratic)5 2039(mechanisms. It)1 709 5 720 2748 t
( open files both to get rid of any open files left by the command)14 2841(recreated shell had to close all its)6 1479 2 720 2868 t
( it had to reopen the special file corre-)8 1700( Then)1 279( redirection.)1 539(just executed and to rescind previous IO)6 1802 4 720 2988 t
( was no)2 372( There)1 322( order to read a new command line.)7 1697(sponding to its terminal, in)4 1270 4 720 3108 t
10 PI f
(/dev)4423 3108 w
10 PA f
(directory)4634 3108 
2000
w
( no memory across commands, because)5 1759(\(because no path names\); moreover, the shell could retain)8 2561 2 720 3228 t
( a further file system convention was)6 1815( Thus)1 302( each command.)2 781(it was reexecuted afresh after)4 1422 4 720 3348 t
(required: each directory had to contain an entry)7 2150 1 720 3468 t
10 PI f
(tty)2900 3468 w
10 PA f
( terminal)1 406(for a special file that referred to the)7 1588 2 3046 3468 t
( by accident one changed into some directory that lacked this)10 2842( If)1 131( opened it.)2 495(of the process that)3 852 4 720 3588 t
( the)1 179( \(Sometimes)1 582( to reboot.)2 479(entry, the shell would loop hopelessly; about the only remedy was)10 3080 4 720 3708 t
(missing link could be made from the other terminal.\))8 2337 1 720 3828 t
( was designed and implemented within a couple of days.)9 2538(Process control in its modern form)5 1532 2 970 3984 t
( system; at the same time it is easy to see how)11 2019(It is astonishing how easily it fitted into the existing)9 2301 2 720 4104 t
( the slightly unusual features of the design are present precisely because they represented)13 3974(some of)1 346 2 720 4224 t
( the separation of the)4 979( good example is)3 787( A)1 141(small, easily-coded changes to what existed.)5 2015 4 720 4344 t
10 PI f
(fork)4679 4344 w
10 PA f
(and)4871 4344 w
10 PI f
(exec)720 4464 w
10 PA f
( a)1 82( most common model for the creation of new processes involves specifying)11 3378(functions. The)1 660 3 920 4464 t
( a forked process continues to run the same program)9 2355(program for the process to execute; in Unix,)7 1965 2 720 4584 t
(as its parent until it performs an explicit)7 1867 1 720 4704 t
10 PI f
(exec)2626 4704 w
10 PA f
( separation of the functions is certainly not)7 1988(. The)1 257 2 2795 4704 t
(unique to Unix, and in fact it was present in the Berkeley time-sharing system [2], which was)16 4320 1 720 4824 t
( in Unix mainly)3 749( it seems reasonable to suppose that it exists)8 2094( Still,)1 267(well-known to Thompson.)2 1210 4 720 4944 t
(because of the ease with which)5 1377 1 720 5064 t
10 PI f
(fork)2125 5064 w
10 PA f
( sys-)1 200( The)1 220(could be implemented without changing much else.)6 2312 3 2308 5064 t
( two\) processes; there was a process table, and the processes)10 2772(tem already handled multiple \(i.e.)4 1548 2 720 5184 t
( initial implementation of)3 1147( The)1 224(were swapped between main memory and the disk.)7 2348 3 720 5304 t
10 PI f
(fork)4471 5304 w
10 PA f
(required)4658 5304 w
(only)720 5424 w
( of the process table)4 873(1\) Expansion)1 715 2 720 5580 t
( current process to the disk swap area, using the)9 2288( of a fork call that copied the)7 1378(2\) Addition)1 654 3 720 5736 t
(already existing swap IO primitives, and made some adjustments to the process table.)12 3793 1 970 5856 t
(In fact, the PDP-7's)3 865 1 970 6012 t
10 PI f
(fork)1868 6012 w
10 PA f
( course, other)2 616( Of)1 170(call required precisely 27 lines of assembly code.)7 2198 3 2056 6012 t
( and user programs were required, and some of them were rather)11 2896(changes in the operating system)4 1424 2 720 6132 t
( a combined)2 552( But)1 211(interesting and unexpected.)2 1244 3 720 6252 t
10 PI f
(fork-exec)2759 6252 w
10 PA f
(would have been considerably more com-)5 1892 1 3148 6252 t
( because)1 388(plicated, if only)2 723 2 720 6372 t
10 PI f
(exec)1872 6372 w
10 PA f
(as such did not exist; its function was already performed, using)10 2958 1 2082 6372 t
(explicit IO, by the shell.)4 1043 1 720 6492 t
(The)970 6648 w
10 PI f
(exit)1175 6648 w
10 PA f
( new copy of the shell \(actually a sort of)9 1880(system call, which previously read in a)6 1797 2 1363 6648 t
(automatic)720 6768 w
10 PI f
(exec)1205 6768 w
10 PA f
( version a process)3 835(but without arguments\), simplified considerably; in the new)7 2788 2 1417 6768 t
(only had to clean out its process table entry, and give up control.)12 2859 1 720 6888 t
(Curiously, the primitives that became)4 1671 1 970 7044 t
10 PI f
(wait)2667 7044 w
10 PA f
( present)1 356(were considerably more general than the)5 1814 2 2870 7044 t
( pair of primitives sent one-word messages between named processes:)9 3100(scheme. A)1 481 2 720 7164 t
cleartomark
showpage
saveobj restore
end
%%PageBoundingBox: 61 60 514 764
%%EndPage: 5 5
%%Page: 6 6
%%PageBoundingBox: (atend)
DpostDict begin
/saveobj save def
mark
6 pagesetup
10 PA f
(- 6 -)2 166 1 2797 480 t
(smes\(pid, message\))1 860 1 1080 900 t
(\(pid, message\) = rmes\(\))3 1035 1 1080 1020 t
( of)1 132(The target process)2 844 2 720 1200 t
10 PI f
(smes)1740 1200 w
10 PA f
(did not need to have any ancestral relationship with the receiver,)10 3061 1 1979 1200 t
( except that)2 517(although the system provided no explicit mechanism for communicating process IDs)10 3803 2 720 1320 t
10 PI f
(fork)720 1440 w
10 PA f
( were not queued; a)4 901( Messages)1 479(returned to each of the parent and child the ID of its relative.)12 2754 3 906 1440 t
(sender delayed until the receiver read the message.)7 2263 1 720 1560 t
( exe-)1 217(The message facility was used as follows: the parent shell, after creating a process to)14 3853 2 970 1716 t
(cute a command, sent a message to the new process by)10 2577 1 720 1836 t
10 PI f
(smes)3338 1836 w
10 PA f
(; when the command terminated)4 1507 1 3533 1836 t
( messages\) the shell's blocked)4 1316(\(assuming it did not try to read any)7 1572 2 720 1956 t
10 PI f
(smes)3634 1956 w
10 PA f
(call returned an error indi-)4 1185 1 3855 1956 t
( the shell's)2 471( Thus)1 274( target process did not exist.)5 1257(cation that the)2 640 4 720 2076 t
10 PI f
(smes)3390 2076 w
10 PA f
(became, in effect, the equivalent)4 1427 1 3613 2076 t
(of)720 2196 w
10 PI f
(wait)833 2196 w
10 PA f
(.)1010 2196 w
( messages,)1 474(A different protocol, which took advantage of more of the generality offered by)12 3596 2 970 2352 t
( initialization)1 594( The)1 225( the shells for each terminal.)5 1280(was used between the initialization program and)6 2221 4 720 2472 t
( for each of the terminals, and then)7 1643(process, whose ID was understood to be 1, created a shell)10 2677 2 720 2592 t
(issued)720 2712 w
10 PI f
(rmes)1033 2712 w
10 PA f
(; each shell, when it read the end of its input file, used)12 2449 1 1228 2712 t
10 PI f
(smes)3708 2712 w
10 PA f
(to send a conventional `I)4 1106 1 3934 2712 t
( for that)2 354(am terminating' message to the initialization process, which recreated a new shell process)12 3966 2 720 2832 t
(terminal.)720 2952 w
( explains why the facility was replaced by the)8 2108( This)1 252( recall no other use of messages.)6 1488(I can)1 222 4 970 3108 t
10 PI f
(wait)720 3228 w
10 PA f
( the present system, which is less general, but more directly applicable to the desired)14 3837(call of)1 273 2 930 3228 t
( also is the evident bug in the mechanism: if a command process)12 3085( relevant)1 407(purpose. Possibly)1 828 3 720 3348 t
( use messages to communicate with other processes, it would disrupt the shell's syn-)13 3751(attempted to)1 569 2 720 3468 t
( a command)2 560( shell depended on sending a message that was never received; if)11 2952(chronization. The)1 808 3 720 3588 t
(executed)720 3708 w
10 PI f
(rmes)1151 3708 w
10 PA f
(, it would receive the shell's phony message, and cause the shell to read another)14 3694 1 1346 3708 t
( had manifested)2 717( a need for general messages)5 1278( If)1 120(input line just as if the command had terminated.)8 2205 4 720 3828 t
(itself, the bug would have been repaired.)6 1809 1 720 3948 t
( some very valuable features)4 1272(At any rate, the new process control scheme instantly rendered)9 2798 2 970 4104 t
( shell as a)3 426(trivial to implement; for example detached processes \(with `&'\) and recursive use of the)13 3894 2 720 4224 t
( sort of special `batch job submission' facility and a)9 2254( systems have to supply some)5 1321(command. Most)1 745 3 720 4344 t
(special command interpreter for files distinct from the one use
2000
d interactively.)10 3416 1 720 4464 t
( very easily indeed, there were some afteref-)7 2005(Although the multiple-process idea slipped in)5 2065 2 970 4620 t
( most memorable of these became evident soon after the new)10 2764( The)1 225(fects that weren't anticipated.)3 1331 3 720 4740 t
( of our jubilation, it was discovered that the)8 1931( the midst)2 442( In)1 142(system came up and apparently worked.)5 1805 4 720 4860 t
10 PI f
(chdir)720 4980 w
10 PA f
( was much reading of)4 1019( There)1 321( directory\) command had stopped working.)5 2018(\(change current)1 712 4 970 4980 t
( anxious introspection about how the addition of)7 2220(code and)1 410 2 720 5100 t
10 PI f
(fork)3384 5100 w
10 PA f
(could have broken the)3 1014 1 3573 5100 t
10 PI f
(chdir)4621 5100 w
10 PA f
(call.)4863 5100 w
(Finally the truth dawned: in the old system)7 1946 1 720 5220 t
10 PI f
(chdir)2696 5220 w
10 PA f
( cur-)1 206(was an ordinary command; it adjusted the)6 1900 2 2934 5220 t
( the new system, the)4 929( Under)1 342( \(unique\) process attached to the terminal.)6 1904(rent directory of the)3 905 4 720 5340 t
10 PI f
(chdir)4832 5340 w
10 PA f
( but this)2 386(command correctly changed the current directory of the process created to execute it,)12 3934 2 720 5460 t
( was necessary)2 667( It)1 124( whatsoever on its parent shell!)5 1413(process promptly terminated and had no effect)6 2116 4 720 5580 t
(to make)1 365 1 720 5700 t
10 PI f
(chdir)1120 5700 w
10 PA f
( turns out that several)4 1004( It)1 126( within the shell.)3 762(a special command, executed internally)4 1785 4 1363 5700 t
(command-like functions have the same property, for example)7 2725 1 720 5820 t
10 PI f
(login)3470 5820 w
10 PA f
(.)3676 5820 w
( system as it had been and the new process control scheme)11 2663(Another mismatch between the)3 1407 2 970 5976 t
( each open file)3 644( the read/write pointer associated with)5 1757( Originally,)1 534(took longer to become evident.)4 1385 4 720 6096 t
( pointer indicates where in the file the)7 1743( \(This)1 283( file.)1 199(was stored within the process that opened the)7 2095 4 720 6216 t
( this organization became evident only)5 1799( problem with)2 664( The)1 235(next read or write will take place.\))6 1622 4 720 6336 t
( a simple command file contains)5 1423( Suppose)1 428(when we tried to use command files.)6 1627 3 720 6456 t
(ls)1080 6636 w
(who)1080 6756 w
(and it is executed as follows:)5 1264 1 720 6936 t
(sh comfile)1 451 1 1080 7116 t
10 S f
(>)1556 7116 w
10 PA f
(output)1611 7116 w
(The sequence of events was)4 1219 1 720 7296 t
cleartomark
showpage
saveobj restore
end
%%PageBoundingBox: 61 46 514 764
%%EndPage: 6 6
%%Page: 7 7
%%PageBoundingBox: (atend)
DpostDict begin
/saveobj save def
mark
7 pagesetup
10 PA f
(- 7 -)2 166 1 2797 480 t
( main shell creates a new process, which opens)8 2066(1\) The)1 417 2 720 840 t
10 PI f
(outfile)3228 840 w
10 PA f
( and)1 195(to receive the standard output)4 1336 2 3509 840 t
(executes the shell recursively.)3 1312 1 970 960 t
( creates another process to execute)5 1529( new shell)2 445(2\) The)1 417 3 720 1116 t
10 PI f
(ls)3137 1116 w
10 PA f
(, which correctly writes on file)5 1343 1 3204 1116 t
10 PI f
(output)4573 1116 w
10 PA f
(and)4871 1116 w
(then terminates.)1 716 1 970 1236 t
( the)1 172( the IO pointer for)4 831( However,)1 495( process is created to execute the next command.)8 2202(3\) Another)1 620 5 720 1392 t
( still 0, because the shell has never written on)9 1991(output is copied from that of the shell, and it is)10 2079 2 970 1512 t
( effect is that the output of)6 1234( The)1 229( processes.)1 483(its output, and IO pointers are associated with)7 2124 4 970 1632 t
10 PI f
(who)970 1752 w
10 PA f
(overwrites and destroys the output of the preceding)7 2306 1 1161 1752 t
10 PI f
(ls)3492 1752 w
10 PA f
(command.)3584 1752 w
( pointers)1 396(Solution of this problem required creation of a new system table to contain the IO)14 3674 2 970 1908 t
(of open files independently of the process in which they were opened.)11 3098 1 720 2028 t
10 PB f
(IO Redirection)1 679 1 720 2268 t
10 PA f
( `)1 66(The very convenient notation for IO redirection, using the)8 2656 2 970 2424 t
10 S f
(>)3692 2424 w
10 PA f
(' and `)2 301 1 3747 2424 t
10 S f
(<)4048 2424 w
10 PA f
(' characters, was not)3 937 1 4103 2424 t
( Like)1 254( but it did appear quite early.)6 1353(present from the very beginning of the PDP-7 Unix system,)9 2713 3 720 2544 t
( has a rather general IO redi-)6 1296( Multics)1 386(much else in Unix, it was inspired by an idea from Multics.)11 2638 3 720 2664 t
( be dynamically redirected to vari-)5 1551(rection mechanism [3] embodying named IO streams that can)8 2769 2 720 2784 t
( version of)2 488( in the)2 296( Even)1 283(ous devices, files, and even through special stream-processing modules.)8 3253 4 720 2904 t
( that switched subsequent)3 1164(Multics we were familiar with a decade ago, there existed a command)11 3156 2 720 3024 t
( and another command to reattach output to)7 2026(output normally destined for the terminal to a file,)8 2294 2 720 3144 t
( under Unix one might say)5 1182( Where)1 344(the terminal.)1 564 3 720 3264 t
(ls)1080 3444 w
10 S f
(>)1176 3444 w
10 PA f
(xx)1231 3444 w
(to get a listing of the names of one's files in)10 1901 1 720 3624 t
10 PI f
(xx)2646 3624 w
10 PA f
(, on Multics the notation was)5 1280 1 2746 3624 t
(iocall attach user)2 744 1 1080 3804 t
10 S f
(_)1824 3804 w
10 PA f
(output file xx)2 594 1 1874 3804 t
(list)1080 3924 w
(iocall attach user)2 744 1 1080 4044 t
10 S f
(_)1824 4044 w
10 PA f
(output syn user)2 697 1 1874 4044 t
10 S f
(_)2571 4044 w
10 PA f
(i/o)2621 4044 w
( sequence was used often during the Multics days, and would have)11 2992(Even though this very clumsy)4 1328 2 720 4224 t
( the Multics shell, the idea did not occur to us or)11 2261(been utterly straightforward to integrate into)5 2059 2 720 4344 t
( the Multics)2 543( speculate that the reason it did not was the sheer size of)12 2606( I)1 94(anyone else at the time.)4 1077 4 720 4464 t
( Labs in Murray Hill, while the shell was)8 1838(project: the implementors of the IO system were at Bell)9 2482 2 720 4584 t
( the shell \(it was)4 727( didn't consider making changes to)5 1563( We)1 200(done at MIT.)2 574 4 720 4704 t
10 PI f
(their)3812 4704 w
10 PA f
(program\); correspond-)1 1011 1 4029 4704 t
(ingly, the keepers of the shell may not even have known of the usefulness, albeit clumsiness, of)16 4320 1 720 4824 t
10 PI f
(iocall)720 4944 w
10 PA f
( [4] lists)2 369( 1969 Multics manual)3 981(. \(The)1 288 3 933 4944 t
10 PI f
(iocall)2610 4944 w
10 PA f
(as an `author-maintained,' that is non-standard,)5 2178 1 2862 4944 t
( its shell were under the exclusive control of)8 2033( both the Unix IO system and)6 1370(command.\) Because)1 917 3 720 5064 t
(Thompson, when the right idea finally surfaced, it was a matter of an hour or so to implement it.)18 4265 1 720 5184 t
10 PB f
(The advent of the PDP-11)4 1166 1 720 5424 t
10 PA f
( by today's standards,)3 979( Primitive)1 464( a going concern.)3 757(By the beginning of 1970, PDP-7 Unix was)7 1870 4 970 5580 t
( of providing a more congenial programming environment than its alterna-)10 3465(it was still capable)3 855 2 720 5700 t
( it was clear that the PDP-7, a machine we didn't even own, was already)14 3420(tives. Nevertheless,)1 900 2 720 5820 t
( 1970 we proposed)3 848( early)1 256( In)1 150(obsolete, and its successors in the same line offered little of interest.)11 3066 4 720 5940 t
( some sense, this proposal)4 1169( In)1 147(acquisition of a PDP-11, which had just been introduced by Digital.)10 3004 3 720 6060 t
( in the series of attempts that had been made throughout the preceding year.)13 3384(was merely the latest)3 936 2 720 6180 t
( the amount of money \(about $65,000\) was an order of)10 2485( First,)1 285(It differed in two important ways.)5 1550 3 720 6300 t
( to)1 113(magnitude less than what we had previously asked;
2000
 second, the charter sought was not merely)14 4207 2 720 6420 t
( to create a system specifically designed)6 1828(write some \(unspecified\) operating system, but instead)6 2492 2 720 6540 t
( The)1 234(for editing and formatting text, what might today be called a `word-processing system.')12 4086 2 720 6660 t
( J. F. Ossanna, who was then and until the end of his)12 2372(impetus for the proposal came mainly from)6 1948 2 720 6780 t
( perhaps too)2 553( our early proposals were too vague, this one was)9 2209( If)1 119(life interested in text processing.)4 1439 4 720 6900 t
( long, however, funds were obtained through the)7 2200( Before)1 339( met with disfavor.)3 850(specific; at first it too)4 931 4 720 7020 t
(efforts of L. E. McMahon and an order for a PDP-11 was placed in May.)14 3158 1 720 7140 t
( at the end of the summer, but the PDP-11 was so new a product that)15 3097(The processor arrived)2 973 2 970 7296 t
cleartomark
showpage
saveobj restore
end
%%PageBoundingBox: 61 46 514 764
%%EndPage: 7 7
%%Page: 8 8
%%PageBoundingBox: (atend)
DpostDict begin
/saveobj save def
mark
8 pagesetup
10 PA f
(- 8 -)2 166 1 2797 480 t
( the meantime, a rudimentary, core-only version of Unix)8 2495( In)1 142( December.)1 498(no disk was available until)4 1185 4 720 840 t
( of the time, the machine sat in a corner,)9 1797( Most)1 278(was written using a cross-assembler on the PDP-7.)7 2245 3 720 960 t
(enumerating all the closed Knight's tours on a 6)8 2114 1 720 1080 t
10 S f
(\264)2834 1080 w
10 PA f
(8 chess board\320a three-month job.)4 1505 1 2889 1080 t
10 PB f
(The first PDP-11 system)3 1095 1 720 1320 t
10 PA f
( internal structure, the first)4 1246( In)1 157( arrived, the system was quickly completed.)6 2033(Once the disk)2 634 4 970 1476 t
( the PDP-11 represented a relatively minor advance over the PDP-7 system;)11 3448(version of Unix for)3 872 2 720 1596 t
( example, there was no multi-programming;)5 1972( For)1 204(writing it was largely a matter of transliteration.)7 2144 3 720 1716 t
( the other hand, there were)5 1296( On)1 207( moment.)1 440(only one user program was present in core at any)9 2377 4 720 1836 t
( interface to the user: the present directory structure, with full path)11 3163(important changes in the)3 1157 2 720 1956 t
( of)1 140(names, was in place, along with the modern form)8 2409 2 720 2076 t
10 PI f
(exec)3321 2076 w
10 PA f
(and)3542 2076 w
10 PI f
(wait)3763 2076 w
10 PA f
(, and conveniences like)3 1100 1 3940 2076 t
( thing about)2 561( the most interesting)3 939( Perhaps)1 419(character-erase and line-kill processing for terminals.)5 2401 4 720 2196 t
( for the system, 8K for)5 977(the enterprise was its small size: there were 24K bytes of core memory \(16K)13 3343 2 720 2316 t
( were limited to 64K bytes.)5 1181( Files)1 254(user programs\), and a disk with 1K blocks \(512K bytes\).)9 2466 3 720 2436 t
( seemed natural, or perhaps)4 1249(At the time of the placement of the order for the PDP-11, it had)13 2821 2 970 2592 t
( the protracted arrival of)4 1125( During)1 380( system dedicated to word processing.)5 1745(expedient, to promise a)3 1070 4 720 2712 t
( usefulness of PDP-7 Unix made it appropriate to justify creating)10 3013(the hardware, the increasing)3 1307 2 720 2832 t
( By)1 174( used in writing the more special-purpose system.)7 2254(PDP-11 Unix as a development tool, to be)7 1892 3 720 2952 t
( had the slightest interest in scrapping)6 1781(the spring of 1971, it was generally agreed that no one)10 2539 2 720 3072 t
( transliterated the)2 799( we)1 165(Unix. Therefore,)1 757 3 720 3192 t
10 PI f
(roff)2474 3192 w
10 PA f
(text formatter into PDP-11 assembler language, start-)6 2394 1 2646 3192 t
(ing from the PDP-7 version that had been transliterated from McIlroy's BCPL version on Multics,)14 4320 1 720 3312 t
( J. Saltzer's)2 489(which had in turn been inspired by)6 1586 2 720 3432 t
10 PI f
(runoff)2824 3432 w
10 PA f
( early summer, editor)3 967( In)1 146(program on CTSS.)2 823 3 3104 3432 t
( by offering to supply a text-)6 1413(and formatter in hand, we felt prepared to fulfill our charter)10 2907 2 720 3552 t
( the time, they)3 650( At)1 167(processing service to the Patent department for preparing patent applications.)9 3503 3 720 3672 t
(were evaluating a commercial system for this purpose; the main advantages we offered \(besides)13 4320 1 720 3792 t
( taking part in an in-house experiment\) were two in number: first, we sup-)13 3453(the dubious one of)3 867 2 720 3912 t
( extended type-box, could print most of the)7 1943(ported Teletype's model 37 terminals, which, with an)7 2377 2 720 4032 t
(math symbols they required; second, we quickly endowed)7 2658 1 720 4152 t
10 PI f
(roff)3414 4152 w
10 PA f
(with the ability to produce line-)5 1451 1 3589 4152 t
(numbered pages, which the Patent Office required and which the other system could not handle.)14 4277 1 720 4272 t
( we supported three typists from the Patent department, who)9 2804(During the last half of 1971,)5 1266 2 970 4428 t
( patent applications, and meanwhile tried to)6 1985(spent the day busily typing, editing, and formatting)7 2335 2 720 4548 t
( hard-)1 268( has a reputation for supplying interesting services on modest)9 2732( Unix)1 267(carry on our own work.)4 1053 4 720 4668 t
( on a machine with)4 861(ware, and this period may mark a high point in the benefit/equipment ratio;)12 3459 2 720 4788 t
( program required care and)4 1266(no memory protection and a single .5 MB disk, every test of a new)13 3054 2 720 4908 t
( the typists)2 518(boldness, because it could easily crash the system, and every few hours' work by)13 3802 2 720 5028 t
(meant pushing out more information onto DECtape, because of the very small disk.)12 3698 1 720 5148 t
( only did the Patent department adopt Unix,)7 1979( Not)1 223( trying but successful.)3 970(The experiment was)2 898 4 970 5304 t
( work, but we achieved)4 1036(and thus become the first of many groups at the Laboratories to ratify our)13 3284 2 720 5424 t
( first PDP 11/45 sys-)4 928(sufficient credibility to convince our own management to acquire one of the)11 3392 2 720 5544 t
( and labored continuously on the)5 1498( have accumulated much hardware since then,)6 2100( We)1 206(tems made.)1 516 4 720 5664 t
( the interesting work has already been published, \(e.g. on the sys-)11 2982(software, but because most of)4 1338 2 720 5784 t
(tem itself [1, 5, 6, 7, 8, 9]\) it seems unnecessary to repeat it here.)14 2781 1 720 5904 t
10 PB f
(Pipes)720 6144 w
10 PA f
( culture of operating systems)4 1314(One of the most widely admired contributions of Unix to the)10 2756 2 970 6300 t
(and command languages is the)4 1416 1 720 6420 t
10 PI f
(pipe)2172 6420 w
10 PA f
( course, the funda-)3 853( Of)1 172( commands.)1 546(, as used in a pipeline of)6 1130 4 2339 6420 t
( the)1 171( Even)1 280( coroutine.)1 479(mental idea was by no means new; the pipeline is merely a specific form of)14 3390 4 720 6540 t
( not unprecedented, although we didn't know it at the time; the `communica-)12 3413(implementation was)1 907 2 720 6660 t
( do,)1 190(tion files' of the Dartmouth Time-Sharing System [10] did very nearly what Unix pipes)13 4130 2 720 6780 t
(though they seem not to have been exploited so fully.)9 2365 1 720 6900 t
( 1972, well after the PDP-11 version of the system was in opera-)12 2895(Pipes appeared in Unix in)4 1175 2 970 7056 t
( a long-time advocate of the non-)6 1481(tion, at the suggestion \(or perhaps insistence\) of M. D. McIlroy,)10 2839 2 720 7176 t
( years before pipes were)4 1297( Some)1 350( coroutines.)1 570(hierarchical control flow that characterizes)4 2103 4 720 7296 t
cleartomark
showpage
saveobj restore
end
%%PageBoundingBox: 61 46 514 764
%%EndPage: 8 8
%%Page: 9 9
%%PageBoundingBox: (atend)
DpostDict begin
/saveobj save def
mark
9 pagesetup
10 PA f
(- 9 -)2 166 1 2797 480 t
( operators, whose left)3 964(implemented, he suggested that commands should be thought of as binary)10 3356 2 720 840 t
( a `copy' utility would be com-)6 1471( Thus)1 291(and r
2000
ight operand specified the input and output files.)8 2558 3 720 960 t
(manded by)1 502 1 720 1080 t
(inputfile copy outputfile)2 1084 1 1080 1260 t
( to sort)2 340( Thus,)1 311(To make a pipeline, command operators could be stacked up.)9 2852 3 720 1440 t
10 PI f
(input,)4264 1440 w
10 PA f
(paginate it)1 487 1 4553 1440 t
(neatly, and print the result off-line, one would write)8 2304 1 720 1560 t
(input sort paginate offprint)3 1210 1 1080 1740 t
(In today's system, this would correspond to)6 1942 1 720 1920 t
(sort input)1 435 1 1080 2100 t
10 S f
(\357)1540 2100 w
10 PA f
(pr)1614 2100 w
10 S f
(\357)1739 2100 w
10 PA f
(opr)1813 2100 w
( immedi-)1 406(The idea, explained one afternoon on a blackboard, intrigued us but failed to ignite any)14 3914 2 720 2280 t
( were several objections to the idea as put: the infix notation seemed too radical)14 3557( There)1 309(ate action.)1 454 3 720 2400 t
(\(we were too accustomed to typing `cp x y' to copy)10 2262 1 720 2520 t
10 PI f
(x)3008 2520 w
10 PA f
(to)3084 2520 w
10 PI f
(y)3198 2520 w
10 PA f
( unable to see how to dis-)6 1137(\); and we were)3 655 2 3248 2520 t
( the one-input one-output)3 1191( Also,)1 295( parameters from the input or output files.)7 1983(tinguish command)1 851 4 720 2640 t
( a failure of imagination!)4 1090( What)1 291(model of command execution seemed too confining.)6 2313 3 720 2760 t
( McIlroy's persistence, pipes were finally installed in the operat-)9 2894(Some time later, thanks to)4 1176 2 970 2916 t
( used the same charac-)4 1001( It)1 118( new notation was introduced.)4 1353(ing system \(a relatively simple job\), and a)7 1848 4 720 3036 t
( example, the pipeline above might have been written)8 2373( For)1 201(ters as for I/O redirection.)4 1167 3 720 3156 t
(sort input)1 435 1 1080 3336 t
10 S f
(>)1540 3336 w
10 PA f
(pr)1595 3336 w
10 S f
(>)1695 3336 w
10 PA f
(opr)1750 3336 w
10 S f
(>)1905 3336 w
10 PA f
(The idea is that following a `)6 1273 1 720 3516 t
10 S f
(>)1993 3516 w
10 PA f
( file, to specify redirection of output to that file, or)10 2249(' may be either a)4 743 2 2048 3516 t
( trailing `)2 402( The)1 217( of the preceding command is directed as input.)8 2107(a command into which the output)5 1511 4 720 3636 t
10 S f
(>)4957 3636 w
10 PA f
(')5012 3636 w
(was needed in the example to specify that the \(nonexistent\) output of)11 3156 1 720 3756 t
10 PI f
(opr)3910 3756 w
10 PA f
( to)1 123(should be directed)2 840 2 4077 3756 t
(the console; otherwise the command)4 1658 1 720 3876 t
10 PI f
(opr)2414 3876 w
10 PA f
(would not have been executed at all; instead a file)9 2290 1 2582 3876 t
10 PI f
(opr)4907 3876 w
10 PA f
(would have been created.)3 1133 1 720 3996 t
( Many)1 313( received, and the term `filter' was soon coined.)8 2121(The new facility was enthusiastically)4 1636 3 970 4152 t
( had imagined)2 650( example, no one)3 770( For)1 208(commands were changed to make them usable in pipelines.)8 2692 4 720 4272 t
(that anyone would want the)4 1254 1 720 4392 t
10 PI f
(sort)2000 4392 w
10 PA f
(or)2181 4392 w
10 PI f
(pr)2302 4392 w
10 PA f
( sort or print its standard input if given no explicit)10 2240(utility to)1 383 2 2417 4392 t
(arguments.)720 4512 w
( annoying was a silly lexical)5 1281( Most)1 283(Soon some problems with the notation became evident.)7 2506 3 970 4668 t
(problem: the string after `)4 1137 1 720 4788 t
10 S f
(>)1857 4788 w
10 PA f
(' was delimited by blanks, so, to give a parameter to)10 2309 1 1912 4788 t
10 PI f
(pr)4247 4788 w
10 PA f
(in the example,)2 678 1 4362 4788 t
(one had to quote:)3 770 1 720 4908 t
(sort input)1 435 1 1080 5088 t
10 S f
(>)1540 5088 w
10 PA f
("pr \2612")1 299 1 1595 5088 t
10 S f
(>)1894 5088 w
10 PA f
(opr)1949 5088 w
10 S f
(>)2104 5088 w
10 PA f
( attempt to give generality, the pipe notation accepted `)9 2514(Second, in)1 463 2 720 5268 t
10 S f
(<)3697 5268 w
10 PA f
(' as an input redirection in a)6 1288 1 3752 5268 t
(way corresponding to `)3 1038 1 720 5388 t
10 S f
(>)1758 5388 w
10 PA f
( could also write, for)4 923( One)1 239('; this meant that the notation was not unique.)8 2065 3 1813 5388 t
(example,)720 5508 w
(opr)1080 5688 w
10 S f
(<)1260 5688 w
10 PA f
(pr)1315 5688 w
10 S f
(<)1415 5688 w
10 PA f
("sort input")1 509 1 1470 5688 t
10 S f
(<)1979 5688 w
10 PA f
(or even)1 330 1 720 5868 t
(pr)1080 6048 w
10 S f
(<)1205 6048 w
10 PA f
("sort input")1 509 1 1260 6048 t
10 S f
(< >)1 135 1 1769 6048 t
10 PA f
(opr)1904 6048 w
10 S f
(>)2059 6048 w
10 PA f
(The pipe notation using `)4 1120 1 720 6228 t
10 S f
(<)1840 6228 w
10 PA f
(' and `)2 283 1 1895 6228 t
10 S f
(>)2178 6228 w
10 PA f
(' survived only a couple of months; it was replaced by the pre-)12 2807 1 2233 6228 t
( the old)2 364( Although)1 492( a pipeline.)2 516(sent one that uses a unique operator to separate components of)10 2948 4 720 6348 t
( course,)1 342( Of)1 165( inner consistency, the new one is certainly superior.)8 2337(notation had a certain charm and)5 1476 4 720 6468 t
( is unabashedly linear, though there are situations in which multiple redi-)11 3265( It)1 119(it too has limitations.)3 936 3 720 6588 t
( compare the out-)3 801( example, what is the best way to)7 1506( For)1 207(rected inputs and outputs are called for.)6 1806 4 720 6708 t
( paral-)1 289( is the appropriate notation for invoking a program with two)10 2724( What)1 294(puts of two programs?)3 1013 4 720 6828 t
(lel output streams?)2 844 1 720 6948 t
( provided a mechanism by)4 1212(I mentioned above in the section on IO redirection that Multics)10 2858 2 970 7104 t
( the)1 179(which IO streams could be directed through processing modules on the way to \(or from\))14 4141 2 720 7224 t
cleartomark
showpage
saveobj restore
end
%%PageBoundingBox: 61 54 514 764
%%EndPage: 9 9
%%Page: 10 10
%%PageBoundingBox: (atend)
DpostDict begin
/saveobj save def
mark
10 pagesetup
10 PA f
(- 10 -)2 216 1 2772 480 t
( it might seem that stream-splicing in Multics was)8 2273( Thus)1 280( sink.)1 244(device or file serving as source or)6 1523 4 720 840 t
( In)1 143( certainly was for its Unix version.)6 1520(the direct precursor of Unix pipes, as Multics IO redirection)9 2657 3 720 960 t
( only were coroutines well-)4 1247( Not)1 232( only in a weak sense.)5 1015(fact I do not think this is true, or is true)10 1826 4 720 1080 t
( as Multics spliceable IO modules required that the mod-)9 2592(known already, but their embodiment)4 1728 2 720 1200 t
( genius of)2 433( The)1 218( for no other purpose.)4 969(ules be specially coded in such a way that they could be used)12 2700 4 720 1320 t
( same commands used con-)4 1283(the Unix pipeline is precisely that it is constructed from the very)11 3037 2 720 1440 t
( this possibility and to invent the nota-)7 1738( mental leap needed to see)5 1188( The)1 221(stantly in simplex fashion.)3 1173 4 720 1560 t
(tion is large indeed.)3 874 1 720 1680 t
10 PB f
(High-level languages)1 974 1 720 1920 t
10 PA f
( and)1 201(Every program for the original PDP-7 Unix system was written in assembly language,)12 3869 2 970 2076 t
( there was no)3 638( Moreover,)1 528( no macros.)2 539(bare assembly language it was\320for example, there were)7 2615 4 720 2196 t
( first interesting lan-)3 944( The)1 233( be complete in itself.)4 998(loader or link-editor, so every program had to)7 2145 4 720 2316 t
( Soon)1 278( version of McClure's TMG [11] that was implemented by McIlroy.)10 3019(guage to appear was a)4 1023 3 720 2436 t
( we could not pretend to offer a real com-)9 1923(after TMG became available, Thompson decided that)6 2397 2 720 2556 t
( I recall, the intent)4 813( As)1 174(puting service without Fortran, so he sat down to write a Fortran in TMG.)13 3333 3 720 2676 t
( of and a com-)4 639( he produced instead was a definition)6 1676( What)1 293(to handle Fortran lasted about a week.)6 1712 4 720 2796 t
( was much influenced by the BCPL language [13]; other influ-)10 2730( B)1 112( the new language B [12].)5 1117(piler for)1 3
2000
61 4 720 2916 t
( syntax, and the very small space into which the com-)10 2472(ences were Thompson's taste for spartan)5 1848 2 720 3036 t
( produced simple interpretive code; although it and the programs it)10 3012( compiler)1 421( The)1 220(piler had to fit.)3 667 4 720 3156 t
( interfaces to the regular sys-)5 1298( Once)1 284( more pleasant.)2 686(produced were rather slow, it made life much)7 2052 4 720 3276 t
( enjoy the benefits of using a reasonable)7 1821(tem calls were made available, we began once again to)9 2499 2 720 3396 t
( write what are usually called `systems programs:' compilers, assemblers, and the)11 3784(language to)1 536 2 720 3516 t
( Multics unreasonable, it was much)5 1567( some might consider the PL/I we used under)8 2054(like. \(Although)1 699 3 720 3636 t
( cross-compiler for the)3 1054( other programs, the PDP-7 B)5 1395( Among)1 405(better than assembly language.\))3 1466 4 720 3756 t
( in B, and in the course of time, the B compiler for the PDP-7 itself was)16 3401(PDP-11 was written)2 919 2 720 3876 t
(transliterated from TMG into B.)4 1403 1 720 3996 t
( fact, a version of the)5 950( In)1 149( to it almost immediately.)4 1159(When the PDP-11 arrived, B was moved)6 1812 4 970 4152 t
(multi-precision `desk calculator' program)3 1850 1 720 4272 t
10 PI f
(dc)2599 4272 w
10 PA f
(was one of the earliest programs to run on the PDP-)10 2321 1 2719 4272 t
( passing thought)2 756( Only)1 280( B did not take over instantly.)6 1355( However,)1 496( before the disk arrived.)4 1087(11, well)1 346 6 720 4392 t
( the same was true of)5 941(was given to rewriting the operating system in B rather than assembler, and)12 3379 2 720 4512 t
( approach was taken)3 945( This)1 251( in assembler.)2 626( the assembler was rewritten)4 1308( Even)1 283(most of the utilities.)3 907 6 720 4632 t
( smaller but still real importance was)6 1657( Of)1 166(mainly because of the slowness of the interpretive code.)8 2497 3 720 4752 t
(the mismatch of the word-oriented B language with the byte-addressed PDP-11.)10 3535 1 720 4872 t
( story of the lan-)4 723( The)1 217(Thus, in 1971, work began on what was to become the C language [14].)13 3130 3 970 5028 t
( C is told elsewhere [15], and need not be repeated)10 2266(guage developments from BCPL through B to)6 2054 2 720 5148 t
( the most important watershed occurred during 1973, when the operating system)11 3685(here. Perhaps)1 635 2 720 5268 t
( that the system assumed its modern form; the most)9 2290( was at this point)4 759( It)1 118(kernel was rewritten in C.)4 1153 4 720 5388 t
( were few externally-)3 997( There)1 326(far-reaching change was the introduction of multi-programming.)6 2997 3 720 5508 t
( more rational and general.)4 1210(visible changes, but the internal structure of the system became much)10 3110 2 720 5628 t
( that C was useful as a nearly universal tool for systems)11 2574(The success of this effort convinced us)6 1746 2 720 5748 t
(programming, instead of just a toy for simple applications.)8 2594 1 720 5868 t
( only important Unix program still written in assembler is the assembler itself;)12 3588(Today, the)1 482 2 970 6024 t
( of the applications programs, although)5 1751(virtually all the utility programs are in C, and so are most)11 2569 2 720 6144 t
( seems certain that much of)5 1242( It)1 124( as well.)2 370(there are sites with many in Fortran, Pascal, and Algol 68)10 2584 4 720 6264 t
( of Unix follows from the readability, modifiability, and portability of its software that)13 3831(the success)1 489 2 720 6384 t
(in turn follows from its expression in high-level languages.)8 2603 1 720 6504 t
10 PB f
(Conclusion)720 6744 w
10 PA f
( old memories is their tendency to take on a rosy glow.)11 2488(One of the comforting things about)5 1582 2 970 6900 t
( the early versions of Unix seems, when described)8 2293(The programming environment provided by)4 2027 2 720 7020 t
( find)1 208( am sure that if forced back to the PDP-7 I would)11 2165( I)1 85(here, to be extremely harsh and primitive.)6 1862 4 720 7140 t
( it did not seem so at the time;)8 1383( Nevertheless,)1 657( and lacking in conveniences.)4 1319(it intolerably limiting)2 961 4 720 7260 t
cleartomark
showpage
saveobj restore
end
%%PageBoundingBox: 61 50 514 764
%%EndPage: 10 10
%%Page: 11 11
%%PageBoundingBox: (atend)
DpostDict begin
/saveobj save def
mark
11 pagesetup
10 PA f
(- 11 -)2 216 1 2772 480 t
( and on the joy of helping to create the)9 1831(the memory fixes on what was good and what lasted,)9 2489 2 720 840 t
( ten years, I hope we can look back with the same mixed)12 2558( In)1 148( better.)1 313(improvements that made life)3 1301 4 720 960 t
(impression of progress combined with continuity.)5 2208 1 720 1080 t
10 PB f
(Acknowledgements)720 1320 w
10 PA f
( grateful to S. P. Morgan, K. Thompson, and M. D. McIlroy for providing early docu-)15 3866(I am)1 204 2 970 1476 t
(ments and digging up recollections.)4 1582 1 720 1596 t
( in describing the evolution of ideas, this paper attributes ideas)10 2787(Because I am most interested)4 1283 2 970 1752 t
( reader will not, on the aver-)6 1305( The)1 225( where it seems most important.)5 1461(and work to individuals only)4 1329 4 720 1872 t
( he reads each occurrence of `we' with unclear antecedent as `Thompson, with)12 3453(age, go far wrong if)4 867 2 720 1992 t
(some assistance from me.')3 1151 1 720 2112 t
10 PB f
(References)720 2352 w
10 PA f
( M. Ritchie and K. Thompson, `The Unix Time-sharing System, C. ACM)11 3293(1. D.)1 352 2 720 2544 t
10 PB f
(17)4402 2544 w
10 PA f
(No. 7 \(July)2 500 1 4540 2544 t
(1974\), pp 365-37.)2 736 1 970 2664 t
( and B. W. Lampson, `SDS 930 Time-sharing System Preliminary Reference)10 3493( P. Deutch)2 491(2. L.)1 336 3 720 2820 t
(Manual,' Doc. 30.10.10, Project Genie, Univ. Cal. at Berkeley \(April 1965\).)10 3225 1 970 2940 t
( system,' Proc. Third Symposium)4 1474( J. Feiertag and E. I. Organick, `The Multics input-output)9 2504(3. R.)1 342 3 720 3096 t
(on Operating Systems Principles, October 18-20, 1971, pp. 35-41.)8 2832 1 970 3216 t
(4.)720 3372 w
10 PI f
(The Multiplexed Information and Computing Service: Programmers' Manual,)7 3178 1 970 3372 t
10 PA f
( of Tech-)2 390(Mass. Inst.)1 474 2 4176 3372 t
(nology, Project MAC, Cambridge MA, \(1969\).)5 2018 1 970 3492 t
( Thompson, `Unix Implementation,' Bell System Tech J.)7 2463(5. K.)1 348 2 720 3648 t
10 PB f
(57)3584 3648 w
10 PA f
( 1978\), pp.)2 461(No. 6, \(July-August)2 867 2 3712 3648 t
(1931-46.)970 3768 w
( of C Programs and the Unix System,' Bell Sys-)9 2127( C. Johnson and D. M. Ritchie, Portability)7 1865(6. S.)1 328 3 720 3924 t
(tem Tech J.)2 488 1 970 4044 t
10 PB f
(57)1508 4044 w
10 PA f
(No. 6, \(July-August 1978\), pp. 2021-48.)5 1697 1 1633 4044 t
( Preparation,' Bell Sys. Tech. J.,)5 1393( `Document)1 547( and J. F. Ossanna.)4 836( W. Kernighan, M. E. Lesk,)5 1208(7. B.)1 336 5 720 4200 t
10 PB f
(57)970 4320 w
10 PA f
(No. 6, pp. 2115-2135.)3 916 1 1095 4320 t
( W. Kernighan and L. L. Cherry, `A System for Typesetting Mathematics,' J. Comm.)13 3984(8. B.)1 336 2 720 4476 t
(Assoc. Comp. Mach.)2 907 1 970 4596 t
10 PB f
(18,)1927 4596 w
10 PA f
(pp. 151-157 \(March 1975\).)3 1131 1 2077 4596 t
( W. Kernighan, `Computer Typesetting of Technical Journals on Unix,')9 3246( E. Lesk and B.)4 704(9. M.)1 370 3 720 4752 t
(Proc. AFIPS NCC)2 780 1 970 4872 t
10 PB f
(46)1775 4872 w
10 PA f
(\(1977\), pp. 879-88.)2 794 1 1900 4872 t
(10.)720 5028 w
10 PI f
( 635 Computer,)2 645(Systems Programmers Manual for the Dartmouth Time Sharing System for the GE)11 3425 2 970 5028 t
10 PA f
(Dartmouth College, Hanover, New Hampshire, 1971.)5 2357 1 970 5148 t
( National Conf. \(1968\),)3 1004( M. McClure, `TMG--A Syntax-Directed Compiler,' Proc 20th ACM)8 2974(11. R.)1 342 3 720 5304 t
(pp. 262-74.)1 478 1 970 5424 t
( Tech. Rep.)2 496( C. Johnson and B. W. Kernighan, `The Programming Language B,' Comp. Sci.)12 3496(12. S.)1 328 3 720 5580 t
(#8, Bell Laboratories, Murray Hill NJ \(1973\).)6 1943 1 970 5700 t
( Richards, `B
3c0
CPL: A Tool for Compiler Writing and Systems Programming,' Proc. AFIPS)12 3950(13. M.)1 370 2 720 5856 t
(SJCC)970 5976 w
10 PB f
(34)1223 5976 w
10 PA f
(\(1969\), pp. 557-66.)2 794 1 1348 5976 t
( and D. M. Ritchie,)4 840( W. Kernighan)2 651(14. B.)1 336 3 720 6132 t
10 PI f
(The C Programming Language,)3 1291 1 2576 6132 t
10 PA f
(Prentice-Hall, Englewood)1 1144 1 3896 6132 t
( Edition, 1979.)2 626( Second)1 369(Cliffs NJ, 1978.)2 653 3 970 6252 t
( Ritchie, S. C. Johnson, and M. E. Lesk, `The C Programming Language,' Bell Sys.)14 3807( M.)1 161(15. D.)1 352 3 720 6408 t
(Tech. J.)1 319 1 970 6528 t
10 PB f
(57)1339 6528 w
10 PA f
(No. 6 \(July-August 1978\) pp. 1991-2019.)5 1747 1 1464 6528 t
cleartomark
showpage
saveobj restore
end
%%PageBoundingBox: 61 123 514 764
%%EndPage: 11 11
%%Trailer
DpostDict begin
done
end
%%Pages: 11
%%DocumentFonts: Palatino-Roman Palatino-Bold Palatino-Italic Times-Roman Symbol
.

0