[HECnet] This is probably been asked already but....

Johnny Billquist bqt at softjar.se
Mon Jul 2 15:13:24 PDT 2012


On 2012-07-02 15:39, Paul_Koning at Dell.com wrote:

On Jul 1, 2012, at 9:02 PM, Steve Davidson wrote:
...
IAS-11 was based on RSX-11D.

I worked on that for my first job at DEC, supporting Typeset-11, which was a newspaper typesetting and advertisement management system.   It was originally implemented as a turnkey product on top of RSX-11D, then ported to IAS.   There was also a Typeset-8, which was created by the group right next to it and used some of the same terminals, but whether anything else carried over I don't know.

The relationship between those two was very obvious, especially since we turned off the timesharing piece and kept only the RSX compatible real time piece of IAS.

Curiously enough, it seems (I never got very close to it) that RSX-11M (and M+) were completely unrelated to -D apart from having a mostly common API (as we didn't call it yet).   RSX-11D very clearly went back directly to RSX-15 -- I once saw a listing of RSX-15 lying around because it was supported very close to where I worked, and a glance at the first few pages showed lots of data structures identical in name, purpose, and layout to what the RSX-11D kernel used.

Paul, you're a wonderful source of information, as usual (both the previous post on DECnet history), and this...

To comment a tiny bit of what I know on RSX. RSX-11M was a clean reimplementation of RSX by Dave Cutler. Allegedly done in a rather short time, with the aim of much better utilization of resources.
As such, it did indeed implement more or less the same API, but the internals are all different. So user applications compile fine on either. The programs might even work fine without recompiling, but I don't know for sure.
Device drivers, and all stuff that knows anything about the kernel is rather different though.

One of the goals of 11M was to get something that could run on a really small PDP-11 without an MMU, which 11M can. (I seriously doubt that could ever be done with -11D.)
-11M+, which came later, was basically reimplementing some of the stuff in -11D, since -11M+ had as the target the large PDP-11 systems. Specifically the 11/70, as well as the never introduced 11/74. Which is why -11M+ also have a very capable online reconfiguration tool (that turned out to be useful in general, but it was specifically written for the 11/74).

So -11M+ requires even more hardware than -11D, but does things differently than -11D.

	Johnny



More information about the Hecnet-list mailing list