Date: Wed, 12 Feb 2014 08:43:43 -0500
From: allison <ajp166 at verizon.net>
To: cctech at
classiccmp.org
Subject: Re: [u-u] Someone's looking for old software
Message-ID: <52FB7A8F.4080306 at verizon.net>
Content-Type: text/plain; charset=ISO-8859-1
> Will VAXELN even run on a 3100 or a 4000? I
thought it was limited to Q-Bus
systems,
but won't be surprised if I'm wrong.
It runs on any VAX. However its not a OS. Its
closer to a IO subsystem
and scheduling
package (RTOS). Like many RTOS packages the basic
system is fairly
useless and not
a user environment. Its used to make real time
systems. Examples that
used it were
LPS40(qbus), LPS20(not Q), LPS32(not Q) and the terminal
software for
the VS2000
(very not Q). If the bus/IO structure is not Qbus or
other DEC standard
bus then IO
is up to the developer.
Allison
VAXELN is/was a host/target setup for developing distributed
realtime VAX applications (ELN = Executive for Local Networks).
Hosted on any sensible VAX/VMS system, with targets being a
subset of VAX, MicroVAX, VAXstation systems. And a few oddities
too (e.g. RTVAX300 VAX-in-a-module, KAV30 RTVAX300-on-VMEbus,
rtVAX1000 (KA630-variant in a MicroVAX2 chassis), etc.
Obscure targets without publically available VAXELN support (?)
include the aforementioned LPS32 and LPS40 VAX-powered printers.
The SPD seems to be hard to find online, as does the VAXELN
Technical Summary, so what follows is from memory. The usual
suspects do seem to have some of the VAXELN docs online in
decw$book format.
The host and target might be separately licenced.
Early VAXELN applications could be written in a multi-threaded
variant of PASCAL called EPASCAL. Later, other standard VAX
languages were supported too: Ada, C, maybe Fortran, the usual
Pascal. Or VAX Macro.
The runtime environment was a dedicated RTOS environment, not VMS
(e.g. no demand paging) although some VMS-compatible facilities
(DECnet, RMS) and routines (some LIB$)were offered.
You compiled your application in the usual way, and linked it with
VAXELN-specific libraries using the usual linker.
The VAXELN system builder was then used to build a bootable system
image. A system image would include the hardware-specific kernel and
any required VAXELN device drivers. Any optional facilities
(VAXELN distributed name service, networking using DECnet and/or TCP,
CLI, Xserver, debug) would be combined with any required user
applications to make a bootable system image file
Booting was supported over the LAN (MOP, maybe bootp), from disk, and
in an odd few cases, from ROM.
A VAXELN source kit was sold at one stage, though it was far from
essential. Pretty much the whole target end was written in PASCAL.
In the early days VAXELN had its own threads-aware debugger (hosted on
VMS and networked to the target).
Once VMS DEBUG got that capability (which came initially with DEC
Ada? sometime around VMS V4?) VAXELN targets could be debugged remotely
using the standard VMS debugger. Deep joy.
Iirc, VAXELN V4.3 or thereabouts used to be on CD in the 1990s (1993?).
There were a couple more versions after that, but no huge changes I can
remember.
When Alpha came along, VAXELN was left behind. DEC worked with Wind River
to get a version of VxWorks for Alpha (VxWorks was an "industry standard"
RTOS, somewhat different in concept to VAXELN). Both claimed to have POSIX
compatibility for the RT stuff. A thin VAXELN veneer was added on top of
VxWorks for those who might be interested.
The product became part of DEC's Embedded+Realtime Group, which was
sold off to SMART Modular Technologies, in the late 1990s (?), and the
product subsequently vanished. I haven't recently been able to establish
the product's current status, though I didn't try real hard.
It was a great RTOS if you wanted to think more about the application
design and less about the intricacies of device drivers and networking and
other such traditional RTOS trivia. I'm told its concepts mapped nicely onto
the "channels and pools" from "Communicating Sequential Processes".
Be interesting to know what lies behind the original enquiry from LinkedIn.
I may be able to offer more info if necessary. It might not be quick, and
the formatting may be messed up again (sorry).
Hope this helps,
regards
John