[HECnet] DECnet/Python --> Tops20 V4.1

Sat Nov 6 01:52:00 PDT 2021

@Thomas & @Paul

Paul helped me in the past with some Rsts Decnet-III stuff and we got that working with Dmc/Dmr, so we got a functioning phase-III router from that and at a point I also introduced the Tops20 V4.1 Decnet problems; I think he got that Simh image from me  for testing, so if one is interested I can provide a Tops20 V4.1 Decnet image.

Problem there was, that it only communicates with other Tops20 4.1 Decnet but notcontrary to expectations with phase-III. This was not supposed to be a working setup as the KS10 was meant to connect with the Kmc/Dup line to  a DN20/200 Pdp11 based network box to join Decnet. These boxes ran Decnet-III on a Rsx11M V3.1 platform and the software load image was generated and built on Tops20 (or Tops10 probably as well) but should be booted from a KL10 based system. As this software was from link libraries, source code was missing.

Simh Pdp11 could be configured to run the network parallel ANF10 without problems, but with some tricks activating the built Decnet image – after all it is RSX11S bootable within the Rsx11M 3.1 environment - on a suitable configured Simh Pdp11 sim, it stops somewhere in the Decnet initialization phase. Too difficult to debug that at the time.

So Tops20 V4.1 Decnet is quite useless as it looks more than Phase-II with enhancements that stops it working with proper phase-III.

It is still a problem to be solved sometime…

Best regards,

Reindert

From: owner-hecnet at Update.UU.SE [mailto:owner-hecnet at Update.UU.SE] On Behalf Of Thomas DeBellis
Sent: Saturday, 06 November, 2021 01:10
To: hecnet at Update.UU.SE
Subject: Re: [HECnet] DECnet/Python

Boy, now that is a pal!

So where would somebody like me be able to get that?  The 4.1 Galaxy has some very interesting things in MOUNTR that were sadly removed that I would like to put in at some point in order to be able to set super-domestic on a per-structure basis.

super-domestic can make your life easier because you have less directory and user numbers overall to deal with.  However, the other side of that coin is that you have less of them total.  If you have a large number of directories, users and structures, you run out of number, maybe, maybe (it's an 18 bit field)  There's also the case of a restored structure with it's own numbers that might clash with super-domestic.  That might happen if you grab a lot of DECUS stuff (although there is brokeness there, too)

Still for, with automated number management, much of the headache (and hence reason) for super-domestic goes away and I'd rather have the extra granularity.

It really depends on how you're going to use the system and how 'Unixy' you want certain things to work.  So I'd rather shut it off for my purposes, but I sure wouldn't want it not to be available.

  _____  

On 11/4/21 7:30 PM, Paul Koning wrote:

A kind soul sent me a TOPS-20 V4.1 SIMH disk a while ago with basic information, and that's what I use for my Phase II testing.  It has in its <DECNET> directory some sources: NETCON bits, NFT, DAP. 

I also realized there's a DECnet-20 V2 (Phase II) user manual on Bitsavers, which combines programming and management documentation.  It mentions that Phase II NCP has a NICE protocol just like the later versions, except that it doesn't seems to be compatible.  At least my NICE listener doesn't like what it hears.  Something else to play with at some point.

            paul

On Nov 4, 2021, at 5:23 PM, Thomas DeBellis <tommytimesharing at gmail.com <mailto:tommytimesharing at gmail.com> > wrote:

Where did you get that NETCON from?  I don't have it.

Or is a Tops-10 NETCON?  (which I wouldn't have, either)

  _____  

On 11/4/21 3:22 PM, Paul Koning wrote:

I found the code, it's in NETCON (specifically, NCP).  So now all I have to do is reverse engineer it.  That's going to be interesting because I haven't looked at Macro-10 in 45 years, and even back then I didn't really know it well at all.

     paul

  _____  

On Oct 28, 2021, at 9:26 AM, Paul Koning  <mailto:paulkoning at comcast.net> <paulkoning at comcast.net> wrote:

I'm guessing it has to do with learning the shape of the Phase II network.  That's an entirely different problem than Phase III and later.  The connect is by name, so object 0, object name TOPOL.  And I don't have anything to answer that request so I have no trace.  I suppose I could build a dummy responder just to see what question is asked.

If indeed it's Phase II topology related, it would make sense for the host not to have that server, and of course it would also go away in Phase III.  The host requests, but does not offer, "intercept" which is node name based routing in Phase II that was implemented only in a few places.  Somewhere I saw that it exists only to get past the front end (on larger machines) which was handled as a separate node so it counts as a network hop.  Without intercept, Phase II only goes a single hop.

I found the code, it's in NETCON (specifically, NCP).  So now all I have to do is reverse engineer it.  That's going to be interesting because I haven't looked at Macro-10 in 45 years, and even back then I didn't really know it well at all.

       paul

  _____  

On Oct 27, 2021, at 9:32 PM, Thomas DeBellis  <mailto:tommytimesharing at gmail.com> <tommytimesharing at gmail.com> wrote:

TOPOL?  Hmm...  No, I hadn't heard of that, either.  It sounds almost familiar, but I don't know why Tops-20 would be asking for it because it  doesn't appear to be serving it, viz:...

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.sonic.net/pipermail/hecnet-list/attachments/20211106/3e2e3b4b/attachment-0001.htm>