THE UNIX NEWSLETTER Volume 5 Number 7 September 1980 CONTENTS Editorilal-:: 5.20 wexscee See biwee das 6 9, Ga: tes wig Wie Bate akese-S5eleveuens aise Fle iotecwsa de bieha sacs eelearees 1 Guidelines for Submission of Newsletter... .scccceccccecececceccecuceuceecvas 1 Letters......... el Diishin Bae ora iDiwhans xe apare Se eis Be woe eiwne See, wal ote ere oi ecu ietere aise eee: Syelelian ser decree 2 Delaware Usenix Meeting Report......c ccc ccc c cece cece cscccccaseccnee Ae ectavenee 5 4.1 What's Happening with UNIX............ Sieueie are Geis weeeraset oriole 6; 0h 89: 91.67 Srananeans a 5 4,2 Kernel Extensions and Performance........cecccercccccceccecsceccccces ase 7 4,3 Languages and Porting C and UNIX... .. cece cece cece eee ceccenrcccecreeves 8 4.4 Vendor Presentations.....cecceeeeeees Si8S Sse 8 OS ose Mee ees ayevers dave un saiereyael aah ‘ 9 WS: Data Bases .rieesss-6 6:6 tas wide sess weit sieveues ws Be, e\ielevere a bo 8 ate Sie ches Shien aatarave 9 4,6 USENIX Business and Overflow ....... ccc cee eect eeeereees hawesergemaseas> 10: 4.7 ARPANET BOF ...........0.-0005 fb iiesailaies Sey biavoraseieusetw'averbie} ar Ba waives é:taiss-epateveraiersrere 10 4.8 Text Processing and Office Automation ........eceeeeeee HS. orc(e ease vee wee. 10 4.9 Communications and Networking ........ccccecesceccececs Blaiasesd Whete eis tee suee 11 DECUS UNIX SIG Progress Report...... ccc ce cc cence esse eects cessccencecesesses 13 ON EX TLADIGS Ss oc.5 tis yet OS wc ole esiale-e Kies ease leie: d brad era oine SiS ee Sis eswlaeteceleten saci: 14 Delaware Conference Tapes ssi csv ceceee cess ewde reves ea edeerewedsaesavceccee 14 UNIX 20. Che: NEWB 25:50. b056's wieredis evereieie sia 4:8 Sed Sw oe el bele eo Siales Sees eedanedeceuasaa: 1 ZILOG Cross-Software for 2-8000 available... ..cec cece eee cercccereccvcesesces 14 Microprocessor cross-assembler from System-Kontakt........-.cccceceeeceuceee 14 Unet communications software from 3 Com..............0.cccccesecuceeee Risidtanaiey, det Xenix to be offered by MiCroSOft.... ccc ce cece cee cee ses caccnccnccecccecces 14 UniFlex offered by Technical Systems Consultants.........ccccececccccccecees 14 Newsletters and Articles about UNIX... ... ccc cece cece ec ceerceccceccetcseccese 14 Books and Reports about UNIX. ...... cece cece cece cs ce rene ccvcceccssccsassecces 14 C in the NewS............ecseee Cee EATNOT SaaS ip acai 's.a("e lela ayaieresarat Stages ese She steer 16 Small-C Compiler from The Code WorkS.....c.cceececccecccecrecvcsccssccsecces 16 C Compiler for 6800 by Wintek......... ajeie wie waraS Sat8/a aisvesece Sis we\er's'S oven elaerecevearey - 10 C Compiler for Data General Systems by Unidot.......cecccecccecccccscccceees 16 Text formatter from Johnson-Laird....... ata die wieieVel atalene dae! Bai sose fa 6 os Si tad ecvecesee 16 Newsletters and Articles about C............ Salo vare oye Goa) vel ahevis. a0. 6e. 9 ele Siei'enew elapse - 16 Books and Reports about C.......eeeceeeees Siete 6 Sibied ere 6 Store wieiers Biecele ere baisieersww es | “16: C Implementation Notes......... oes ec seanen Cee eee wees eee sacs o Sin date ie-davela ve wee 17 Notice This newsletter may contain information covered by one or more licenses, copy- rights, and non-disclosure agreements. Permission to copy without fee all or part of this material is granted to Institutional Members of the Usenix Associa— tion provided that copies are made for internal use at the member campus or plant site. * UNIX is a trademark of Bell Laboratories Editorial This issue of j;login: concen- trates more on technical material than Usenix Association matters. This is an emphasis which I expect to continue and expand. The departmental organi- zation of the newsletter introduced in this issue will be expanded as required to accommodate material sub- mitted by you and other readers. The section on UNIX and UNIX-derived sys- tems will concentrate on issues related to the UNIX environment including software systems which run on UNIX based computer systems. In contrast the section on C will be devoted to discussions of C language related issues both under UNIX and under other systems such as CP/M. As part of the effort to discover the extent of the acceptance of C, an implementor's coupon is being pub- lished. Anyone having an implementa- tion of C is invited to return the filled-out coupon. We would like to use this newsletter to discuss c related issues just as Pascal News covers Pascal related issues. Finally, let me comment on the plans for numbers 3, 4, and 5 of ylogin:. A copy of the Delaware Usenix meeting Proceedings will be mailed to each ;login: subscriber as issue number 5 and dated June, 1980. Numbers 3 and 4 will contain some pre- viously unpublished meeting minutes and some correspondence which has not been published. Numbers 3 and 4 will be edited by Mel Ferentz at Rockefeller University. Guidelines for Submission of Newsletter Material I would like to use the modern text preparation and communications facilities of UNIX to as great an extent as feasible in the preparation of the Newsletter. I have established an account on our PWB/UNIX system so that those who are able to provide us with machine manageable material can do so. The telephone number is (512) 474-5511. The login name is login and the password is usenix. (The system is also host utexas-11 on the ARPANET.) For those submitting paper copy of material, please produce your copy on a daisy wheel printer or similar high quality printing device. Line printer produced copy is typically not adequate for reproduction. Copy should be on 8 1/2" by 11" paper with a 1" margin on left, right, and bottom and 1 1/2" margin on top. U. S. Mail submissions should be addressed to: Login Newsletter Computation Center The University of Texas Austin, Tx 78712 Attn: Wally Wedel September 1980 2 slogin: Letters USENIX ASSOCIATION Box 8 Rockefeller University 1230 York Avenue New York, N.Y. 10021 13 October 1980 Wally Wedel Editor slogin: Computation Center The University of Texas Austin, TX 78712 Dear Wally: In the August issue of ;login: you published a letter regarding the issue of employment notices and recruitment activities at USENIX meetings. My letter of March 31 to the membership, along with Mr. Fuchs' letter was sent to ;login: prior to the Delaware meeting so that members of the Association would be informed of a resolution adopted unanimously by the Board of Directors which could affect activities at the meeting. When we were unable to produce the newsletter in time, we made copies of the letters which we posted and dis- tributed in Delaware. I feel that it would be in the best interests of the voting members to pub- lish these letters now, so that those who read ;login: who were not present in Delaware will be fully informed. I enclose these letters for inclusion in the September issue. Congratulations on getting the production of ;login: rolling again! Very truly yours, Lou Katz, President September 1980 3 login: Wie Chitty | University Computer Canter 555 West 57 Street, New York, NY 10019 Telephone 212 977-8000 Or GW Work | Office of the Director | March 26, 1980 | ! Dr. Louis Katz President- USENIX Department of Pharmacology College of Physicians and Surgeons | 630 West 168th St. | New York, NY 10032 Dear Lou: Having learned that USENIX is considering whether or not to permit personnel recruiting at USENIX meetings and/or personnel related advertising in the USENIX newsletter, I am writing to you to express our feelings as a member of the organization. I believe that as long as USENIX remains an or- ganization whose voting members are institutions as op- posed to individuals it would not be in the best in- terests of those organizations to permit either of the aforementioned activities. Most institutions are finding it increasingly difficult to attract, and especially to retain competent technical personnel. Our institution budgets funds to send people to user conferences like USENIX to share information with other installations. We would be very reluctant to send our people to a confer- ence where one of the sanctioned activities was personnel recruitment. Although it is clear that some institutions (and certainly individuals) would gain by allowing recruiting at the meetings, I believe that if institutions are to be encouraged to send their best people to USENIX meetings then they should be able to do so without assisting them to find new jobs. Therefore I feel that personnel re- cruiting should be officially proscribed at USENIX spon- sored meetings. lf maximum uncensored distribution of the newsletter is desired then I believe that personnel ad- vertising should be viewed in a similar way. Sincerely yours, Se Ira H. Fuchs Director September 1980 4 slogin: USENIX ASSOCIATION 80x 8 THE ROCKEFELLER UNIVERSITY 1230 YORK AVENUE NEW YORK, N. Y. 10021 March 31, 1980 To the Members : As a result of our session at the USENIX meeting in Boulder, I have given much thought to the issues of employment registries, notices of jobs available, jobs wanted, etc. as activities of the USENIX Asso- ciation. As all the members of the Board discussed these issues ex- tensively, it became clear that the USENIX Association in respect to these activities should be compared to SHARE or GUIDE rather than to ACM or IEEE. ; It then. was clear to us that the recruiting/job seeking activities are inappropriate for the USENIX Association. Mr. Ira Fuchs, Director of the CUNY Computer Center wrote a letter to me at my suggestion which states this case better than I might. This letter is reproduced below. As a result, of these discussions the Board of Directors unanimously approved a resolution prohibiting recruiting and posting of employment notices at meetings of the association or in the newletter, :Login; So. Site Lou Katz President LK: 4g September 1980 Delaware Usenix Meeting Report 4.1 What's Happening with UNIX This session was chaired by Dan Grim of the University of Delaware Comput— ing Center. The session opened promptly at 9 a.m. on 17 June 1980. 4.1.1 Welcome to Delaware Dr. Pete Warter, Chairman of the Electrical Engineering Department at the University of Delaware welcomed us to the campus. 4.1.2 More Gnus from DEC -- Bill Mun- son, Digital Fourteen months ago Bill's group for UNIX support and visibility within DEC was established as part of the DEC Telco group. Currently there are 16 people working in this area within DEC. There are 6-9 people working on Unix, 4-5 working on compilers, and a couple of people working on software tools. The compiler group is currently working on a C compiler for the VAX. They are also concentrating on several other areas. 1. Improved DEC Field Support for UNIX sites, DEC has entered into an arrangement with Western whereby DEC can supply a version of UNIX to Field Service Groups responsible for the maintenance of UNIX systems. The idea is for DEC to have a DEC supported, known baseline UNIX system which can be booted on a customer machine and allowed to run diag- nostics to exercise the system. License coverage has been obtained for 6th Edition, 7th Edition and UNIX/32. No PWB license was obtained. This ver- sion of UNIX should be available to Field Service the Fall. people during 2. Understanding UNIX performance on DEC systems. Bill's group has started a careful study of UNIX performance on PDP-11 systems. The object is to understand 3 login: thoroughly just how UNIX utilizes the hardware in the hopes of pro- ducing hardware which will run well under UNIX. They are con- centrating on instrumenting an 11/70 version of UNIX initially. They will then produce a Simula— 67 queueing model to explore several different areas. The next step will be to look care- fully at UNIX on VAX hardware. Bill said they put UNIX up on an 11/44 with changes to 23 lines of code. He noted that the 11/60, RLO1 disks, and KMC communica- tions gear had been particularly troublesome areas for UNIX users in the past. DEC hopes to avoid problems with new equipment by virtue of the effort of his group. 3. Introduction of more UNIX flavor into DEC systems. Another area in which the group is working is to make available more UNIX capa- bilities in DEC systems particu— larly VAX/VMS. Among the items they are studying are the possi- bility of introducing byte stream files into VMS. The record oriented file system of VMS seems to be particularly onerous to UNIX users used to the simplicity of UNIX files. Another area is the development of an ability to run multiple processes from a terminal. 4.1.3 DECUS SIG Progress Report -- Wally Wedel. I reported for Mark Bartelt on the progress of the UNIX SIG being formed within DECUS. A DECUS board meeting was held last week to consider the formation of the SIG. Results aren't yet in. Mark raised several questions which I presented to the group. @ What should be the relationship between DECUS SIG and Usenix? Discussion in the halls after the session produced at least the following sentiments. September 1980 1, Many Usenix people enjoy the relative vendor independence represented by the Usenix Association. The appearance of alternative alternative systems like Onyx adds much flavor to the Usenix meet-— ings. 2. DECUS is too big. 3. DECUS meetings are too expensive. 4, The possibility of combining a DECUS trip with a Usenix trip was appealing to several people. e What function can the SIG serve as ameans of communicating with the hardware and software planners at Digital? Many people seemed to think that this func- tion would be among the most valuable that the DECUS SIG could provide. @ How do we insure a good relation- ship with Digital? There was some concern over the fact that many UNIX users tend to take a rather dim view of DEC software. The elegance and simplicity of UNIX systems contrasts rather strongly with the complexity of most DEC software. Many people pointed out however that there are strengths in several pieces of DEC software and DEC needs to be encouraged to improve on those strengths. e@ What special problems are people encountering making UNIX co-exist with DEC operating systems? There are a significant number of sites which run either a DEC sys— tem or UNIX depending on the application required. Many are encountering strange problems in going from one system to another. The SIG appears to be the place to deal with these. ;login: @ How do we bring Software Tools progress to the attention of DECUS and DIGITAL? It would appear that some DECUS sessions presenting the Software Tools work would be beneficial to all concerned. Mark is also looking for DECUS ses- sions for November. Among the topics he wants to cover are 1. A UNIX Introductory Overview. 2. A UNIX Applications Panel. 3. A UNIX "wizards" session. Mark's address is: Mark Bartelt Caltech 356-48 Pasadena, CA 91125 (213) 795-6811 x 2663 4,1.4 UNIX Extensions at USGS — Peter Ward, USGS Peter Ward described work being done to develop a network of UNIX iabora- tory systems within the Earthquake Labs of the U, S. Geological Survey. The USGS has 7 11/70s and 7 11/3438 scattered across the U. S. They are using UNIX in a scientific environment for the analysis of geological data. They make heavy use of the UCB work on UNIX. Fortran-77 is the primary language being used. The compiler is producing object code which is too big. It also has insufficient error reporting and diagnosis. This group has IMSL available for Fortran-77 on UNIX, They have developed a package called Geolab as an interactive language to work with seismic data analysis. They have boot tapes for 11/23s through 11/70s. They use the Haley-Joy text overlay extensions heavily. These mods have allowed them to put £77, the ex editor, and adb on an 11/34. 4.1.5 ISO Pascal from VU ~- Gary Fos- tel, Intermetrics Gary Fostel announced the availability of a new version of Andrew September 1980 Tannenbaum's VU Pascal which conforms to the proposed ISO standard. It can be obtained for $50 from Intermetrics. 4.1.6 Tektronix and UNIX -- Clement Cole, Tektronix Tektronix is using UNIX in a software development environment. Most of the people in this environment have found both the Cybers and the DEC-10 inade- quate in one way or another. A series of tools have been developed for the UNIX 7th Edition environment. They have encountered some problems with V7. The most serious problems were with the terminal handler. Init tended to hang when sent a kill -2. “S/*Q support was added to the system. They also added some other features such as “O and echoing of several other eontrol characters. Shell changes include adding a logout com- mand and a 5 minute mail check. They use the Rand MH mail system. Mike Wahrman of Rand is now the distribu- tor. They use the CMU bulletin board system. Tektronix is heavily commit- ted to Pascal. They use the VU-Pascal package but have done much work on it. Most of the work was converting it to use standard i/o. 4.1.7 Equation macro package -- Mel- vin Lax, CCNY Dr. Lax reported on a troff macro package he has developed for working with equations. It is on the confer- ence distribution tape. 4.1.8 BRLNET Progress Report -- Mike Muuss, BRL Mike Muuss reported on the development progress of a closely coupled local network being developed at the Ballis-— tics Research Laboratory at Aberdeen, Maryland. 4.1.9 Another Flavor of UNIX —- Bill Plauger, Whitesmiths Bill Plauger reported on the Whi- tesmiths IDRIS UNIX look alike system. IDRIS uses a heap to allocate kernel tables. It has about 6000 lines of slogin: resident code 10% of which is bler. assem- They have concluded that the 8 bit micros are below 'C' level for IDRIS. They are porting to the 68000. 300 lines of code are devoted to memory management. Another 500 lines of code is being changed to move IDRIS. 4,2 Kernel Extensions and Performance This session was chaired by Ian John- stone of Bell Labs. 4.2.1 Extensions to PWB to support MASCOT ~- David Tilbrook, BNSR David Tilbrook of Bell Northern Software Research presented his MASCOT system. This is an implementation of the British Army developed MASCOT sys- tem. BNSR is using it to develop micro based systems for Office Automa- tion and Switching applications. 4.2.2 Cross Linking UNIX Terminals -- Jeff Schriebman This is a set of mods to the kernel to allow one user to monitor another user's terminal session, It was developed for LLL and is in the public domain. 4.2.3 Berkeley 7th Edition Perfor- mance Improvements -- Ed Gould, UCB Ed reported that a 1.5 megabyte 11/70 system serving 40 users under 6th Edi- tion went to a 25 user system under Tth Edition. The biggest performance change they made was to use 1024 byte disk blocks instead of 512 byte blocks. They also moved the i/o buffers out of kernel space. They allocate 100 buffers. They use a hashed lookup based on block number for the buffers and for the inode table. They also reduced the number of direct block pointers from 8 to 4 in the inode. The Ken Thompson disk quota stuff was also installed. Bob Kridle (esvax.kridle@berkeley) con- trols distribution. September 1960 4.2.4 Delaware 7th Edition Improve- ments ~- Ron Reiser, UDEL UDEL moved i/o buffers, proc table and elists out of ispace. They hashed all these tables. They implemented proc queues. Their scheduler schedules once a second now. Disk was config- ured to put swap space in middle. They are also plugging one disk drive into the RH70 which is also the tape controller. They only use the tape drive for backups. 4.2.5 Maryland Improvements -- Bob Kirby, University of Maryland Bob has rewritten the tape driver at Maryland for reliability and more flexibility. He has installed a writ- able I space mode to the kernel for LISP compiler support. 4.2.6 UNIX Guru Panel 4.3 Languages and Porting C and UNIX Joe Yao of Science chaired this session, Applications 4.3.1 News from Western Electric ~- 1 Arms Al announced that he would be leaving his present duty in two months. His replacement will be Otis Williams. The Ritchie patent on set user id has been dedicated. There are no longer any patents covering UNIX for which fees must be paid Western. Al also announced that a letter is being mailed to all licensees stating that C compilers are fully interchangeable between systems. Bell Labs has decided not to release the Equipment Test Package. This package requires support from an internal error logging version of UNIX which is not scheduled for release. In the educational licensing area, the agreement has been reduced to two pages to minimize costs to Western. Individual cpu agreements can go out over Al Arms’ signature. Mr. Baldwin's signature is no longer required. 7th Edition manuais may be obtained from Irma Biren for $120 per Set, There is no proprietary slogin: information in the manuals. UNIX/370 is now available within AT&T. It is a joint IBM/AT&T product which runs under TSS. A Parasite system is being considered for release. It uses an LSI as a satellite processor to a reg- ular UNIX. Al asked for expressions of interest in Kernighan's troff work and UNIX release 4 work from the UNIX community. A new version of C/370 is being considered for release. 4.3.2 UNIX on the UNIVAC 1100 ~- George | Ronkin, Bell Lab Labs, Pis- cataway Dr. Ronkin discussed an implementation of UNIX for the UNIVAC 1100 computer series. It runs under UNIVAC's EXEC 8 operating system and provides a Unix interface to the user. The implemen- tation required four man-months to do the C compiler, two man-years to do the kernel, and two man-years to do the utilities. 4.3.3 7th Edition UNIX on the 28000 -- Craig Forney, Onyx Craig described the development of UNIX for the Onyx C8000 system. They made the conversion in about 6 months. Interdata/P-E UNIX ~~ Paul Cub- bage, Wollongong Group 4.3.4 Mr. Cubbage described the implementa- tion of 7th Edition UNIX for the Perkin-Elmer 7/32, 8/32, 3220 and 3240 computer systems. They can run 30 users on a 3220 with 750k bytes of memory. Three man-years to do the port. Richard Miller of Wollongong did the port. 4.3.5 The C Machine -- Carl Howe, BBN Carl described progress on the BBN MBB machine now called the C/70. They have no floating point on it. Other— wise it runs a 7th Edition UNIX rea- sonably well. The MBB (microprogramm- able building block) has a writeable eontrol-store that can be tailored to almost any application. The C/70 microcode has been developed to pro~ vide a C instruction set. The machine September 1980 is intended to run only C code and not any assembly language. The hardware characteristics of the MBB are 16K x 32 bit microcode store, 20 bit word length, 1 megabyte unmapped address space, 135 nanosecond cycle time, and 1024 fast registers. There are so many registers that the compiler doesn't save registers between func- tion calls. It just uses new regis— ters. There are a few problems with this approach involving context switches but the microcode has been written to minimize the time to save the registers during a context switch. Tne C/70 instruction set is still being developed but most of it is com- plete. The function linkage is very fast, being three times faster for a function call than an 11/70. The machine provides direct access to the hardware from C, making assembly language programs unnecessary. The instruction set supports 10 bit, 20 bit, and 40 bit arithmetic. There are 18 addressing modes available includ- ing register, local, extern, array, and struct addressing modes. Forty opeodes can use all addressing modes while there are 43 specialized opcodes that cannot use all these addressing modes, The microcode memory has plenty of room for adding additional instructions if needed. The C/70 assembler and compiler have been run- ning for one year and the microcode has been up and running for six months. 28000 -- Jean Butler, Boeing Aerospace Ms. Butler's group is using UCB developed software to develop code for 28000 computer systems on a VAX machine running UNIX. During the question and answer period someone stood up and announced that the Mitre Corporation has a portable C compiler for the Z8000 that they are willing to give away. ; login: 4.3.7 Portable C, Pascal, and F77 -- Randall Howard, Mark Williams Co. This new company is developing a UNIX look-alike system called COHERENT. Tney are developing a full set of utilities in addition to the languages named above. 4.3.8 Maryland ULISP -- Bob Kirby, Maryland University Bob described his LISP for UNIX. implementation 4.4 Vendor Presentations This session was chaired by Ed Szur- kowski of University of Delaware. 4.4.1 HCR Products and Services -- Mike Tilson Mike described HCR's software pro- ducts. They include a _ superset of ANSI BASIC, and a nice RT-11 Emulator. 4.4.2 Onyx Presentation -- Dennis Ferland, Onyx Systems Dennis gave an excellent presentation reflecting on the place of UNIX in the mini and micro computer world. He also presented an execellent overview of the Onyx System. 4.4.3 The Coherent Operating System —- Robert Schwartz, Mark Willi- ams Co. Robert gave a brief review of his company's plans for a UNIX-like operating system and utilities. They are doing everything from scratch using the 7th Edition UNIX manuals as their bible. 4.5 Data Bases This session was chaired by Ed Adami of the National Bureau of Standards, September 1980 4.5.1 Navy use of INGRES —- Neil Groundwater, Analytic Discip- lines Neil described work done by Analytic Disciplines using UNIX and INGRES to develop a system for tracking computer hardware reliability. 4.5.2 The MRS System -- John Korna- towski, University of Toronto John described the MRS data base sys- tem and presented plans for its suc~ cessor system. 4.5.3 Publications, Indexes, and Bibliographies -- Mike Lesk, Bell Labs Dr. Lesk described work he has done using inverted index utilities under UNIX to work with bibliographic data. 4.5.4 UNIX Museum Data Base -~ Richard Wit Witt, , Field M Museum Richard has been working as a_ consul- tant to the Field Museum in Chicago in developing an artifact collection and recording system. 4,6 USENIX Business and Overflow This session was chaired by Lou Katz, USENIX Association President. 4.6.1 MMDF: An Experiment in memo distribution —- David Farber, Univ of Delaware Dr. Farber presented an overview of their MMDF system. 4.6.2 Usenix Business -- Lou Katz Lou presented a brief state of Usenix report. 4.6.3 3D Graphics Editor ~- Mike Muuss, BRL Mike gave a very polished of the Graphics system others have developed for BRL. They have some flavor of VG hardware and have developed an interesting graphics presentation which he and 10 ; login: editor. He also had a movie which displayed some of the capabilities of the editor, Design -- Bruce Borden, Rand Bruce presented an overview of the design of the Rand MH mail system. 4.6.5 A Xerographic Printer -- Kurt Akeley, University of Delaware Kurt presented a deSign overview of a printer they have developed at the EE Department at University of Delaware. It uses bit mapped memory to define the primitives (or characters) to be printed. This device also was func- tioning during the open house. Dr. Warter, who used to be an engineer for Xerox, is very enthusiastic about production prospects for this printer. 4.7 ARPANET BOF Mike O'Dell who is now at LBL chaired this informal discussion of ARPANET support problems in the UNIX commun- ity. There is lots of concern over the long leader orders issued last January saying that only long leaders would be supported after 1 January 81. Doug Shannon of NRL has a quick and somewhat dirty solution to the problem which Greg Noel at NOSC is cleaning up. Ken Shroder at Lincoln Labs has long leader support running under 7th Edition UNIX. The DEC IMP11-B which is a KMC based IMP interface was described as a loser by a couple of people because of timing problems. 4.8 Text Processing and Office Auto- mation Dennis Mumaugh of NSA chaired this session. 4.8.1 OPUS, Office Automation System -- Bill Maldnee, RLG Assoc. Bill described the OPUS system which RLG has built around UNIX. September 1980 4.8.2 CAT Phototypesetter Fundamen- tals -- Dave Buscher, JHU Dave gave an interesting and informa-— tive presentation on the CAT photo-— typesetter. He also volunteered to be an information repository on photo- typesetter support in UNIX. His address is: Dave Buscher JHU/APL Johns Hopkins University Laurel, MD 20810 (301) 953-7100 x 7307 4.8.3 Physical Review Production Status -- Stu Kern, rn, Physical Review Stu reported on their producing the experience on Physical Review using the UNIX system. They are currently producing about 8,000 pages per year using UNIX and will go to 30,000 pages per year next year. 4.8.4 Text Processing Training -- Mary Ann Jackson Mary Ann has prepared some training manuals for Word Processing people using UNIX. She made a _ strong plea for a better screen editor under UNIX. 4.8.5 Document Retrieval and Produc-— tion -- Ben Domenico, NCAR Ben's presentation was in the nature of a needs list for an integrated Document management system. He volun- teered to serve as a focal point for people developing tools which might meet these needs. His address is: Ben Domenico NCAR P.O. Box 3000 Boulder, Colorado 80309 4.8.6 E, New Rand Editor and LA — David Yost, Rand Following up on Bill Plauger's earlier remark that talking about editors was like talking about religion, Dave 1 s;login: preached the gospel of the Rand edi- tor. He has rewritten the editor, adding new features. He has also developed a C function package called LA to manipulate line oriented files. Both of these systems will be avail- able this Fall. Mike Wahrman at Rand will handle distribution. 4.8.7 Graphics and Text -- George Toth, JHU George presented a progress report on a system he is developing to integrate figures and other graphical material into text. 4.8.8 Classification by Paragraph -- Dennis Mumaugh, DOD Dennis reported that new Government security regulations are requiring that documents be given security clas- sifications on a paragraph by para- graph basis. 4.8.9 Text Processing Tools at AFDSC -- == Walt Lazear, AFDSC Walt described the UNIX systems which have been developed at AFDSC in sup— port of text processing needs of several Air Force installations. 4.8.10 OFS, Office Forms System —- John Kornatowski, Univ of Toronto John described the OFS system now being released by the CSRG at Univer- sity of Toronto. 4.8.11 Text Analysis Tools -- Lorinda Cherry, Bell Labs MH Dr. Cherry gave a very interesting but too brief review of work she has done in developing some programs called Style and Diction to analyze the rea- dability of textual material. The programs are available to UNIX sites and run under 7th Edition UNIX. 4.9 Communications and Networking Doug Shannon of NRL chaired this sion. ses—- September 1980 4.9.1 UNIX/ARPANET history —- Steve Holmgren, Mitre Steve gave an excellent presentation of the history of the development of the UNIX ARPANET software. 4.9.2 A Simple Solution to the Arpanet Long Leader Situation -- Doug Shannon, NRL Doug presented a quick and dirty solu- tion for converting to 96 bit leaders. They map the 96 bit ARPANET headers into 32 bit headers in their NCP. The only change required is in the NCP and it is transparent to users and other programs. They map high numbered hosts into low numbered hosts that they have decided they don't ever want to talk to. The implementation required about two weeks. Greg Noll (greg@nrl) has also done this for their Dec-10. 4.9.3 Introduction and Overview of UUCP —- Steve . Bellovin, Trian- gle Universities ‘es Computation _ Center overview to UUCP and talked about some of the things that they use it for at TUCC. They are part of USENET and UUCP gives them an inexpensive way to link into a net- work. Steve presented an Duke University Duke started USENET and introduced it at the last Usenix conference. Any Unix site can be part of USENET as long as they have a dial in port. An auto dialer is not necessary. Duke has arranged with some sites to call them on regular intervals to do file and mail transfers. 4.9.5 Unix - Z8000 Local Networking -- Steve Holmgren, Mitre Corp. Steve described their use in a local network. of Z8000's 12 ; login: 4.9.6 DATAKIT Network at Bell Labs ~~ Greg Chesson, Bell Labs Dr. Chesson gave a very interesting review of Networking within the Bell Labs environment. September 1980 Here's the latest news (mostly good!) about the state of the proposed UNIX special interest group ..... Although nothing in the way of a writ- ten confirmation has been received as yet, I spoke with Frank Anderson by phone recently, and he reports that the DECUS executive board has given its approval to the existence of the SIG as described in our original application for recognition. They have not yet, however, approved the name “UNIX SIG" as the name of the group. There are a number of reasons for this somewhat peculiar turn of events, associated primarily with the worry that (1) most SIGs which already exist are oriented toward some DEC product, which (2) is something which Digital supports, and therefore (3) the lead that tal. existence of a "UNIX SIG" might some persons to infer erroneously UNIX is being supported by Digi- What we've been asked to do, there- fore, is to present the board with a list of alternative names for the SIG, and presumably the board will approve one or more of them. Frank Anderson mentioned that during the discussion the name "Software Tools SIG" was sug- gested and rejected. This seems sen- sible, inasmuch as the board recog- nizes that although this group will certainly have an interest in imple- mentations of "Software Tools" goodies on various other operating systems, the main thrust of the group will nonetheless be UNIX, Frank also men- tioned that there was a possibility that the board may, after having looked at the alternatives, finally approve "UNIX SIG" as our name after all. A few other suggestions for a name have included "Friendly Operating Sys- tems SIG" (which Digital might con- Sider a bit insulting, since it car- ries with it an implicit implication that DEC operating systems aren't friendly!) and "Special Software and 13 slogin: Operating Systems SIG" -- This name is actually stolen from the VAX SIG's Special Software working group, which has been involved with many of the things that the UNIX SIG would tend to be interested in (although limited to the VAX), plus other non-DEC VAX operating systems. This will probably be the last thing you'll receive by mail until the SIG starts publishing a newsletter (after the Fall symposium, most likely). It's my understanding that our mailing list can be forwarded to the DECUS office, so that those of you who are DECUS members will automatically become members of the SIG. If you don't wish to have your name included on the list sent to DECUS, please let me know. If you're not a DECUS member, this may be the last you'll ever hear from the SIG, so think about joining! Mark Bartelt Caltech 356-48 Pasadena, California 91125 (213) 795-6811, ext 2663 P.S. There have been several submis- sions for UNIX-related sessions at the fall symposium. See the symposium schedule for details. September 1980 UNIX Tidbits Delaware Conference Tapes Ed Szurkowski reports that the University of Delaware has distributed all of the Delaware Usenix Distribu- tion tapes. Anyone who thinks they should have gotten one and did not, should contact: Ed Szurkowski Department of Electrical Engineering University of Delaware Newark, Delaware 19711 (302) 738-8005 If you call, please do so after 2 p.m. Eastern time. UNIX in the News: ZILOG Cross-Software for Z2-8000 avail- Zilog, Inc, of Cupertino, Cali- fornia is marketing a cross-software package that runs under 7th edition UNIX, Cost for source is $10,000 for first CPU and $3,000 for each addi- tional CPU. A binary only license is $3,000 for the first CPU and $1,000 for each additional CPU. The software runs on PDP~11/44, PDP 11/45 and PDP 11/70 systems. Microprocessor cross-assembler from System-Kontakt. System-Kontakt of Bedford, Mas— sachusetts is marketing a line of microprocessor cross-assemblers for UNIX. Prices range from $2,000 to $3,500. Unet communications software from 3 3 Com Corp of Menlo Park, Cali- fornia has announced a communications slogin: software package for 7th edition UNIX. The cost of a right-to-use source license is $7,300 for the first CPU and $4,300 for each additional CPU. Microsoft of Bellevue, Washington is offering an adaptation of 7th edi- tion UNIX called Xenix for 16-bit microprocessors. Costs are to range from $2,000 to $9,000 depending on the number of users. UniFlex offered by Technical Systems Consultants. Technical Systems Consultants of West Lafayette, Indiana is offering an Operating system for a Motorola 6809 based micro computer system. The sys- tem is described as "very UNIX-like". Cost is $450 per copy. Newsletters and Articles about UNIX InfoWorld of October 13, 1980 carries an article entitled "The Com- ing of UNIX", Computer Magazine for August, 1980 is a special issue on Translator Writing Systems. Articles by Al Aho, Stephen Johnson and Susan Graham all treat different aspects of the topic. UNIX is mentioned prominently in all three articles. Books and Reports about UNIX "A 32-Bit Processor Design", Stephen C. Johnson, Computing Science Techni- cal Report #80, Bell Laboratories, Murray Hill, Nd, 07974. ABSTRACT This paper describes a user-level instruction set for a 32-bit processor. The September 1980 machine is simple, compact, and well suited to the C language. In fact, even for 16-bit applications it is reasonable to expect the instruction space for C pro- grams to be 10 to 15% smaller than for the PDP-11. 15 slogin: performs with about 95% accuracy. It works at a rate of about 340 words per second. The programs may be used in assigning stress to synthesized English speech, where it is useful, and often essential, to have the word classes of the words in the text. For example, the "A Simple But Realistic Model of words perfect, record, Floating-Point Computation", W. Ss. transfer, and contrast have Brown, Computing Science Technical different patterns of stress Report #83, Bell Laboratories, Murray if used as nouns than they Hill, NJ, O7974. have if they are used as verbs. The programs may also be useful in readabil- ity studies and, of course, in parsing English. ABSTRACT This paper presents a model of floating-point com- putation, intended as a basis for efficient portable mathematical software, The model involves only simple familiar concepts, expressed in a small set of environ- ment Parameters. Using these, both a program and its documentation can tailor themselves to the host com- puter. "PARTS - A System for Assigning Word Classes to English Text", L.L. Cherry, Computing Science Technical Report #81, Bell Laboratories, Murray Hills, NJ, 07974 ABSTRACT PARTS is a set of pro- grams to assign word classes (parts of speech) to English text by rule. The system works on the theory that if word classes can be assigned or partially assigned to some of the words in a sen- tence, the word classes of the words in the rest of the sentence can be deduced by using some general rules of sentence structure and word order. The system has been tested on a sample of about 10,000 words of text and September 1980 C in the News The Code Works of Goleta, Cali- fornia is offering a Small-C compiler which runs under CP/M and _ produces code for CP/M. Cost is $15. The Wintek Corp of Lafayette, Indiana is offering a C compiler which runs on 6800 systems and produces code for the 6800. Cost of the compiler is $495. C Compiler for Data General Systems by Unidot. Unidot, in Sunnyvale, California markets a C compiler for Data General AOS and RDOS systems. Cost is $6,000 for AOS and $5,300 for RDOS, Text formatter from Johnson-Laird. Johnson-Laird, Inc. of Portland, Oregon is marketing a text formatter for Sanders Technology Media 12/7 printers. The formatter is written in C and runs on MP/M or CP/M with at least 48k bytes of RAM. Price is $175. Newsletter and Articles about C Computer Business News of 22 Sep-— tember 1980 carries a letter about the C Language. Computer Business News! issues of 22 August 1980, 1 September 1980, and 8 September 1980 featured articles on Pascal versus c as programming languages. Books and Reports about C "Classes: An Abstract Data Type 16 ;login: Facility for the C Language", Bjarne Stroustrup, Computing Science Techni- cal Report #84, Bell Laboratories, Murray Hill, NJ, 07974. ABSTRACT Language constructs for definition and use of abstract data types ease the design and maintenance of large programs. This paper describes the C class con- cept, an extension to the C language providing such con- structs. A class is defined using standard C data types and functions, and it can itself be used as a building block for new classes. A class provides a way of res— tricting access to a data structure to a specific set of functions associated with it, without incurring signi- ficant overheads at compile time or at run time. September 1980 17 slogin: C Implementation Notes Implementor/Maintainer/Distributor /*Give a person, address and phone number*/ Machine/System Configuration /*Give both the machine on which the target code runs and the machine on which the compiler runs*/ Distribution /#*Who to ask, how it comes, option, and at what price*/ Documentation /*What is available and where*/ Maintenance /#Is it maintained?#/ Standard /*How does it measure up to Reference Manual in "The C Programming Language" #/ Measurements /*Speed and Space for compiler and compiled code*/ Reliability /*Any information about field use or sites*/ Development Method /*How was it developed and what is the source language? Underlying character set?#/ Library Support /*Is standard I/O library provided? Any other libraries?#/ September 1980 18 j;login: (fold here) Place Postage Here Login Newsletter Computation Center The University of Texas Austin, Texas 78712 U.S.A. (fold here) Note: Login publishes all the checklists it gets. Imple- mentors should send us their checklists so C users can judge their implementations for their merit. Otherwise we must rely on rumors. Please feel free to use additional sheets of paper. The Usenix Association PUR POSE: The Usenix Association is an organization of Western Electric licensees and sub-licensees formed for the purpose of exchanging information and ideas about the UNIX operating system and the C Progamming Language. MEMBERSHIP: Four classes of membership in Usenix are offered: 1. Institutional Membership. Institutional Members are the voting members of the Usenix Association. This class of membership is open only to licensees or sub-licensees of Western Electric Co. 2. Non-voting Institutional Membership. This class of membership is open to corporate affiliates of AT&T. 3. Individual Membership. Open to employees of class 1 and 2 members and others who are bound by the software agreements with Western Electric and its licensees. 4, Public Membership. Open to anyone with a bona fide interest in the purpose of the Usenix Associ- ation. Facts about UNIX and the Programming Language C The UNIX operating system was developed by Ken Thompson and Dennis Ritchie of Bell Laboratories in Murray Hill, N.H., during the early 1970's. The C Programming Language was developed originally by Thomp- son and Ritchie as the implementation language for UNIX. UNIX is made available to the public by Western Electric Co. through its patent licensing office in Greensboro, North Carolina. A fine overview of UNIX and C was published in the Bell System Techni- eal Journal, Vol. 57, No.6; Part 2, in August 1978. The C Programming Language is described in the book The C Programming Language by Brian Kernighan and Dennis Ritchie published in 1978 by Prentice Hall. Login Newsletter V5n7 Computation Center The University of Texas at Austin Austin, Texas 78712