VCF East XIII -- Bill Degnan -- Introduction to VAX/VMS

[@lawrencedoliveiro9104]

12:19 One difference between VMS and Unix-type systems is that, under VMS, you did most of your work in a single process. This process was created when you logged in, and your shell (the standard VMS one was called“DCL”) was mapped into this process in a privileged mode. When you used a command to run a program, that was mapped into the same process in ordinary user mode, and when it completed running, it was deleted from the process address space and control reverted to DCL. So the process was reused for running multiple programs until it was terminated when you logged out. It was also possible to spawn subprocesses and detached processes to run programs, but this was not considered a routine way of doing things. Contrast this with Unix-type systems, where the shell runs in its own process and has no special privileges, and most commands involve the spawning of entirely separate processes to run the relevant programs. To make this work, process creation has to be a relatively cheap operation on Linux and other Unix-type systems, because it happens so frequently. It was expensive on VMS, and something to be avoided as far as possible. Also consider that Microsoft Windows NT (including consumer versions of Windows from XP onwards) very much derive from the VMS heritage in this and other ways.

[@jimmykjazz]

COBOL on VAX/VMS was great ! My favorite feature was the ability to dynamically generate a table in memory that other programs can access. It was especially useful when monitoring various remote devices which would be done with an assembly program and then the COBOL was well suited to access the internal table and display the status to users. Loved it !

[@vcfederation]

Thank you for sharing your experience! :-)

[@lawrencedoliveiro9104]

9:29 Bit of trivia: the precursor to the X11 GUI protocol that was widely adopted on *nix systems was named X10, and that code was written to run only on VAXstations. X11 represented the evolution of this towards hardware independence.

[@tommeggison1466]

The greatest operating system ever written hands down.

[@vcfederation]

It was truly an impressive OS!

[@paultraynorbsc627]

Indeed it was

[@kd1s]

Oh in 1993 I managed a Data General Eclipse runing AOS/VS II. Functionally it was the equivalent to VMS. And one Unix//Linux systems to change the login root pw just boot in single user mode.

[@vcfederation]

Interesting! Thanks for sharing!

[@UVOS-z4q]

Wow very good speaker and topic! I loved this!

[@lawrencedoliveiro9104]

53:11 SIMH is a whole collection of emulators for classic machines from a whole range of vendors. But its principal developer is (seems to be?) Bob Supnik, who used to be a DEC engineer and senior executive.

[@TrolleyMC]

jesus christ dude how many comments are you gonna fucking leave?

[@StephenGutknecht]

I was pondering today near the end of February 2022, that Vax/VMS in the late 1980's as an OS and API seems better than Windows 11 or Ubuntu 20.04 Linux, or even OpenBSD. Sure, you would have to modernize it, but the command line and editing environment just seemed better organized.

[@vcfederation]

Yeah. Some of the early OS were really well designed. That's why they are still used by some places because they simply work!

[@lawrencedoliveiro9104]

46:11 I don’t recall VMS even had the concept of disk partitioning. But it had the opposite concept, in a way, of a “volume set”, where multiple disks could appear as a single filesystem volume.

[@lawrencedoliveiro9104]

33:54 DEC did play a significant part in the early adoption of Ethernet. The original Xerox system only had a bandwidth of 3 megabits/second; it was DEC that insisted that 10Mb/s would be more “future-proof”, even if it did raise the cost.

[@lawrencedoliveiro9104]

4:32 Or perhaps byte-addressability was looking to be more and more important, and the PDP-10, as elegant and popular as it was, was a 36-bit architecture which did not have that capability. The byte-addressable PDP-11 had become wildly successful, but its 16-bit address space was clearly becoming an onerous constraint. VAX was very definitely intended as a spiritual successor to the PDP-11: an all-new 32-bit architecture from the ground up, rather than trying to tack on extensions to the PDP-11 instruction set (thank goodness), but nevertheless the name stood for “Virtual Address eXtension”, the “extension” in question being to the original PDP-11 concept.

[@eb999uj]

Its still the most finetunable permission concept I ever managed, the privilege system is fantastic.

[@vcfederation]

Pretty cool, huh?

[@eb999uj]

@@vcfederation I used to be operator on VMS-Clusters, remember and love it to this day

[@MrTherende]

If I remember correctly, $255$ (or $1$, $2, etc.) was used to identify the disk and tape devices on a cluster member, also known as the SCSid. It's a little hazy now, I haven't worked on a VMS system in close to 30 years; I booted my first VAX (11/780) in March 1980 at Kodak. A typo for us hard core dec'ies, DecNet should be written as DECnet

[@vcfederation]

There are a lot of people that are still heavily involved with this system. Lots of hobbyists are into DEC stuff.

[@mercster]

When I went to community college in 1995/1996, our servers were VAXen running OpenVMS. It's what our programming classes were run on (VAX assembler, etc). Our email was on VMS. Our shell accounts were on VMS (I still hate you DCL.) By that time I was already a Linux kid. But these systems are still fascinating to me.

[@mercster]

In the early-mid 90s we called computers boxen, called VAX machines VAXen, called UNIX machines UNIXen... I dunno why you have a problem with this nomenclature. :-)

[@mercster]

I'm 44 years old and was using the Internet and server OSes in the early 90s... I dunno how old you are.

[@mercster]

Much later I worked with DEC UNIX on Alphas, Sun systems running SunOS/Solaris, HP PA-RISC boxen running HP-UX, and SGI boxen running IRIX (my least favorite.)

[@mercster]

Oh I also admin'd some IBM AIX systems.

[@mercster]

PDP-10? VAX was the successor to the PDP-11. OK you're right, maybe you shouldn't say "VAXen" ;-) Good talk anyway.

[@lawrencedoliveiro9104]

57:51 “MCR” was originally a command for RSX-11 compatibility. But it became established as a shorthand for running executables in the SYS$SYSTEM directory, i.e. MCR «prog» is a shorter way of writing RUN SYS$SYSTEM:«prog» (“RUN” can be abbreviated to just “R”, if you like). Here’s another difference from Unix/Linux-type systems: commands in the DCL shell were not simply the names of programs to run. Some of them would run external commands, others would not. So to run an arbitrary given program, you needed to explicitly use the RUN command (or this weird “MCR” shortcut, as above).

[@craig893]

btw: MCR was an acronym for Monitor Console Routine

[@MrTherende]

mcr authorize was my favorite. MCR: Monitor Console Routine

[@KD5NJR]

VAX was a little before my time . I got to college to find some Sun workstations , but PCs / Windows were taking over in about all fields . Thanks a lot Windows :)

[@awuma]

You forget about Linux...

[@alicewyan]

Great talk! Are the slides available for download?

[@vcfederation]

Contact him through his website: www.kennettclassic.com/

[@kellingc]

I'm sorry, this camera operator is horriable. I'm getting sea sick.