VenturCom's VENIX
VenturCom's VENIX
Mixing REAL Time With REAL UNIX Is Not Magic...It Is Technology.
The Company
According to the Massachusetts Secretary of State, VenturCom was founded on November 11th, 1980. The Articles of Organization list the founders as Michael Corwin, Jeffrey B. Gertz, Douglas R. Mook, and Myron L. Zimmerman.
As I was looking over the Articles of Organization, one thing stood out to me: the listed purpose of the corporation. That section is 13 paragraphs long. Most of it is extremely broad and could apply to any business venture. (There is even a section that says, “To loan to any person, firm or corporation any of its surplus funds, either with or without security.”)
The first paragraph is the one that we should be most interested in. It states, “To engage in the business of providing computer and computer system related services; including, but not limited to selling or leasing time shared computer services, providing local and remote computer-based data processing services, consulting and advisory services, program design, the marketing of computer programs, the education and training of consultants, analysts and programmers, computer programming, installation, purchase, sales, repair, manufacture, and distribution thereof.”
Besides Venix, VenturCom also worked with Windows NT and Windows CE, especially embedded and real-time applications. They worked with Microsoft to get some of their tools included in the Windows CE Embedded Toolkit.
One of their longest-running product lines is their RTX products. RTX is short for Real-time Extension and allows users to create real-time applications.
In early 2005, the company changed their name to Ardence. The following year, Citrix Systems acquired the company. In 2008, former Ardence executives bought the Ardence software back and created IntervalZero.
The Product
Venix was first demoed at COMDEX/Spring ‘83. It was one of the first UNIX systems for IBM PCs. It was based on Version 7 Unix from AT&T with some enhancements from BSD, such as the “vi screen editor”. Version 2 added support for DOS/Fat partitions. Version 3 was based on UNIX System V, AT&T’s first commercial Unix. The last version of Venix to be released was 4.2 in 1994.
While researching this article, I came across an interesting letter to the editor in the June 1985 issue of Unix World:
VENDOR GETS FREE AD
Dear Editor:
This letter represents yet another in my unceasing crusade to assure that publications accurately portray Venix when it is referenced. Venix/86, developed by VenturCom and published by Unisource Software Corp. of Cambridge, Mass., was the first licensed Unix operating system for the IBM PC/AT, PC/XT, and certain compatible personal computers. In your February issue, C. A. Felong and Harry Avant reviewed Xenix from The Santa Cruz Operation. In the article they incorrectly report and misrepresent characteristics of Venix on several occasions, namely:
(one) They state that “Venix requires a little less disk space. ”Well, yes, 3.5 Mbytes is less than the 6-9 Mbytes that Xenix reportedly needs. I, however, would consider 42 percent to 61 percent less to be more than “a little less.” Since this difference represents at least one-fourth of the PC/XT’s hard disk. I’d call that significant.
(two) They state that “Venix is slightly less expensive.” Xenix with the development package and text processing is $1350. Venix/86 is either $800 (two-user license) or $1000 (three-to-eight user license). Thus Venix/86 is either 26 percent or 41 percent cheaper. That too I consider to be a significant rather than “slight” difference.
(three) Venix/86, Version 2.0 (Encore), released in November 1984, has record locking.
(four) I do not agree that Xenix performance is “within 5 percent of Venix. ” Perhaps the fact that Xenix requires much more ram to operate (for example, a minimum 384K bytes versus 256K bytes) was not taken into consideration in the comparison. 1 note that a full 640K bytes was used in your evaluation of Xenix!
Thanks as always for the opportunity to look over your shoulder.
Sincerely,
Clayton D. Clatur
Executive Vice President
Unisource Software Corp.Editor’s Reply: What are you and Greg Graham trying to hide? Over the last year, various editors from this magazine have attempted to shake loose a review copy of Venix 186 from you or VenturCom to prove or disprove what you say about Venix 186, but to no avail. It’s time for Unisource and VenturCom to stop stonewalling, and either put up or shut up. —Philip J. Gill
(Note: VenturCom created Venix and Unisource Software sold it. Source: PC Tech Journal, May 1985 “UniSource is the exclusive licensed distributor of VentureCom’s VENIX/86.”)
Mark S. Zachman wrote a review of VENIX for the June 12, 1984 issue of PC Magazine entitled “A Venerable UNIX”. He introduced the operating system by stating, “IBM recently garnered attention when it announced its version of UNIX for the IBM PC, PC/IX. But VenturCom, Inc., of Cambridge, Massachusetts, quietly introduced an implementation of UNIX Version 7 for the PC a year ago. In fact, I saw a working version of VenturCom’s VENIX/86 on a PC at COMDEX in May 1983 in Atlanta.
At the time, VENIX had a few bugs, but it was functional. VenturCom has managed to sell quite a few copies to people looking for a more professional operating system than PC-DOS. VENIX also has a major selling point that PC/IX lacks: It supports up to two additional users connected to the XT serial ports.
VENIX is based on UNIX Version 7, so it differs slightly from UNIX System III and System V, the two newer major releases by AT&T. System III comes with the Source Code Control System (SCCS), a way for programmers to keep track of different versions of a program. System V, which may become the AT&T standard, is a major upgrade with communications between simultaneous tasks and improved file locking. According to Gig Graham, vice-president of VenturCom, the company plans to support System V completely. He did not mention a release date for a new version of VENIX, however.
Philosophically, VENIX is conservatively designed. Rather than talking directly to the IBM PC, it processes all system functions through the BIOS. The advantage of such an approach is that VENIX may be easily transported to a number of XT compatibles, but the penalty you pay for portability is a slight loss in speed, particularly in the display. Disk operations, however, seem to run as fast as with PC/IX.”
He finished the review with the following comments:
“I like the overall approach of VENIX/86. Some VENIX-specific features that improve task communications, such as semaphoring, have been implemented in UNIX System V, which indicates their utility.
I was a bit disappointed that a number of the less useful functions were missing or not working. I suppose that VenturCom’s finite resources were applied to implementing important features and improving the documentation. What VENIX attempts, for the most part, it accomplishes.”
The December 1984 issue of Dr. Dobb’s Journal contained a long article entitled “Varieties of Unix: An Introductory Guide to Microcomputer Unixes” by Alan Walworth. The article had this to say about Venix: “VenturCom’s VENIX is an IBM PC implementation of Version 7 Unix with Berkeley enhancements, including vi and the C shell. Unlike PC/IX, VENIX can handle up to three simultaneous users. Awk, lex, and yacc are included, but lex is apparently incomplete. Troff is absent. Simple graphics functions are provided for tasks such as drawing lines and circles. On-line reference materials, available on fullblown Unix systems via the man command, are not included in VENIX. A BASIC interpreter is provided. Semaphores are available for interprocess communication, a step in the direction of System V’s IPC capability. The Final Word is supplied as an optional word processing alternative. Of special interest to people dreaming of portable Unix systems is the plan (perhaps a reality by the time this is published) to offer a version of VENIX as an option on Data General’s new 10-pound PC.”
In January 1985, Mark G. Sobell wrote about Venis 2.0 in his “Industry Insider” column for Unix Review:
Back on the desktop front, I talked about Venix 2.0 with Gig Graham, Executive Vice President of VenturCom. The biggest ques tions raised in my mind by VenturCom’s efforts to put UNIX (or Venix, VenturCom’s port of UNIX) on a desktop machine without a hard disk were: “Will it run fast enough to be usable?” and “How do you manage to fit all the UNIX utilities and the user’s data on the system?”
The latest release of Venix helps alleviate the speed problem with sticky memory, a capability akin to the UNIX sticky bit that keeps a program from getting swapped off the swap disk. Sticky memory keeps a program in memory as long as another program doesn’t need the space. It is implemented as a fifo list of programs that always tries to hold the most recently run programs in memory.
As far as fitting UNIX onto a floppy disk, Graham maintains that the kernel and most individual utilities are actually quite small. (However, the entire Venix system ships on 10 floppies.) VenturCom’s solution to the problem of fitting UNIX on a floppy-based computer is to provide the user with a system disk that includes a subset of the utilities — the user has full use of the second disk for data and application programs. This approach will be satisfactory in environments where the user only needs to use as many programs as will fit on the system disk (including the kernel, swap area, device drivers, and system files). It will be tedious, however, to use the system with more programs because, unlike MS-DOS, you cannot just pop a disk out and insert a new one. With a UNIX system, you must unmount the disk, remove it, insert the new disk, and finally mount the new disk before using it.
Augie Hansen wrote an in-depth review of Venix for the February 1986 issue of PC Tech Journal entitled “VENIX: Setting UNIX Standards”. In it, he notes:
VenturCom has brought to market the first UNIX System-V compatible operating system that runs on the PC/XT and PC/AT. VENIX is a port of AT&T sources to the PC family and to several other microcomputer and minicomputer products. Along the way, some features have been added to permit access to DOS programs and data files; a sprinkling of specialty programs has been included as well. Other System V release 2 products from the Santa Cruz Operation and IBM (via Microsoft and Interactive Systems) have been announced or will be shortly.
So why all the interest in System V? Mostly this is a matter of setting standards to avert long-term disarray in the UNIX community. Viewed from a distance, UNIX has taken on a Hydra-like appearance because of efforts by third parties and AT&T itself to make UNIX do things its designers never intended.
Offshoots of the seventh edition (called version 7), with many Berkeley extensions, are hot in scientific and CAD/CAM circles because of additions made in graphics support, among other things. Other UNIX vendors have added helpful realtime features, usually at the expense of multitasking performance. Still others have emphasized modifications to the file system to facilitate database applications.
…
VENEX System V is a full-featured package that offers fast operation. Presently it is flawed by several minor bugs and a few major ones. However, VenturCom’s track record of support for earlier VENEX versions gives encouragement that these bugs will be found and fixed. Updates will be made available to licensed owners as the improvements are made; the prices for updates have not been established.
The DOS interface is the area most in need of significant improvement in this system. A developer of custom applications, for example, depends on a flexible and reliable connection be tween UNIX and DOS. The VENEX/DOS interface is neither flexible, nor reliable at this point. Nevertheless, this is a com plete package that, when the bugs are fixed, will be a good value in a multiuser, multitasking operating system for IBM XT and AT class machines. In the meantime, this is a positive step toward a common UNIX.
Venix Booting on a Rainbow 100 from the Internet Archive
Finally, Micro/Systems Journal published a review of VENIX System V in their May/June 1987 issue. The reviewer (Jim Rosenberg) concluded, “My final verdict is an enthusiastic thumbs sideways. I trust VenturCom to produce the finest UNIX kernel money can buy on PC-compatible equipment, but the poor position of VENIX in the marketplace as far as third-party product support, the lack of a large-memory model C compiler, and SCO’s DOS cross-libraries make it SCO’s ballgame to lose. VENIX works extremely well. If you don’t need third-party software, and especially if you have some reason to mistrust KENIX, VENIX is an excellent piece of work and as “nothing but UNIX” is outstanding. If there are skeletons in the SCO closet, of course, all bets are off.”
Would you like to try out Venix? Here are some resources to help you do that:
Virtually Fun wrote a series of articles on Vernix/86 and how to run it using MESSS/MAME and PCem.
Warner Losh’s talk at FOSDEM ‘22 on a Venix/86 emulator
Did you ever use VenturCom's VENIX? Do you know anything about its history? Let us know in the comments below.
What computer ads would you like to see in the future? Please comment below. If you enjoyed it, please share it with your friends and relatives. Thank you.
