HECnet

hecnet@lists.dfupdate.se

15556 discussions

by Johnny Billquist

I have been planning for a long time to post a small reflection when I reached 1000 nodes registered in the hecnet nodename database. This happened tonight. I think it is a pretty cool thing. There are now 1000 nodenames registered in this small hobby DECnet. I guess you could say it's actually not that small. However, I know that rather few machines are actually online, and it might even be that the majority of nodes registered have never been online. My guess is that maybe 5% of the registered machines are usually online. But I do not have any hard data to back this up. HECnet started out about 20 years ago from a desire I had to hook up a PDP-11 I had at home, to some machines at my university, using DECnet. At the time, I didn't have any TCP/IP for RSX, and the only way to get any kind of networking was to try and come up with some way of getting DECnet up. My first implementation basically just forwarded a serial port communication between two Unix machines. And on each end I then hooked that serial port into a PDP-11 running RSX, and used DDCMP for the actual DECnet link. This worked, but was obviously not that fast, as the serial ports were limited to 9600 bps. After a year or so, I figured I could instead write a small program that would forward ethernet packets. Using UDP I basically had the same property as a local ethernet, but it could be located somewhere pretty far away. From a DECnet point of view, it would appear as if they were on the same ethernet segment, while in reality they were nowhere near. As DECnet have pretty long timeouts on things, it turned out this worked without a hitch, and I could achieve much better throughput. Not long after that, the first other users were hooked up to HECnet as well. This was maybe around 2003 or so. Another data point is that there are 113 different persons that have one or more nodenames registered. Happy milestone, everyone, and thanks for being around. Johnny -- Johnny Billquist || "I'm on a bus || on a psychedelic trip email: bqt(a)softjar.se || Reading murder books pdp is alive! || tryin' to stay hip" - B. Idol

2 years

NETPTH Messages

by Thomas DeBellis

Now that I'm done with Kermit-20, I wanted to get back to finishing up some other work that I had been doing. FAL, DAP and NFT are in good enough shape for me to leave them alone for the moment, so that's another rabbit hole I've gotten myself out of. I'm now looking fix an error that I somehow introduced into MMAILR (SMTP over DECnet). Fixing this required modifying DDT, which I had to do to fix another issue I had introduced into Kermit. I wanted to have a quick look at what Tops-20 hosts might be up on HECnet and wrote up a little batch job to use NETPTH to test connectivity. NETPTH is a nifty utility to find paths between DECnet nodes. It builds a connection the NCU's of various nodes to determine the path that a message might take to get there. Think of it like traceroute or ping -R (RECORD_ROUTE option). It almost never works... Before I roll up my sleeves and jump into this I was wondering if anybody had any observations about the errors I'm seeing? ------------------------------------------------------------------------ *TINA*::*PAMINA*::*KLIO*::*TWLGHT*::*RARITY*::*FLUSHY*::*RBDASH*::*APPLEJ*::*PINKIE*::*DISCRD*::*TWENEX*:: From Via Back Thru Cost Hops VENTI2 (2.522) =>NI-0-0 / BRG-0 <= A2RTR (2.1023) -1 -1 A2RTR (2.1023) =>DMC-31 / ?NETCFO Can't find Output Circuit parameter *SAMT20*:: ?NETURN Unreachable node *TOMMYT*:: From Via Back Thru Cost Hops VENTI2 (2.522) =>NI-0-0 / NI-0-0 <= TOMMYT (2.520) 1 1 *BITXT2*::*BITXT0*::*MINDY*::*FALLON*::*JOSHUA*::*OLAF*:: From Via Back Thru Cost Hops VENTI2 (2.522) =>NI-0-0 / BRG-0 <= A2RTR (2.1023) -1 -1 A2RTR (2.1023) => ?NETCFO Can't find Output Circuit parameter *WALACH*:: From Via Back Thru Cost Hops VENTI2 (2.522) =>NI-0-0 / BRG-0 <= A2RTR (2.1023) -1 -1 A2RTR (2.1023) =>DMC-29-206 / ?NETCFO Can't find Output Circuit parameter *SOL*:: From Via Back Thru Cost Hops VENTI2 (2.522) =>NI-0-0 / BRG-0 <= A2RTR (2.1023) -1 -1 A2RTR (2.1023) =>MUL-59-1016 / DMC-1 <= STORTR (59.1016) -1 -1 STORTR (59.1016) =>GRE-16 / ?NETCFO Can't find Output Circuit parameter *FENCER*:: From Via Back Thru Cost Hops VENTI2 (2.522) =>NI-0-0 / BRG-0 <= A2RTR (2.1023) -1 -1 A2RTR (2.1023) =>MUL-59-1016 / DMC-1 <= STORTR (59.1016) -1 -1 STORTR (59.1016) =>GRE-16 / ?NETPTH NTMAN failed, reason: Parameter missing $

2 years

OpenVMS E9.2-1/x86

by Zane Healy

There is now an OpenVMS/x86 system on HECnet, as I would expect/hope, DECnet Phase IV seems to be working just fine. It’s routing through my VAX/VMS 5.5-2 system. Zane

2 years

Tops-20 Kermit-20 v5.3(230)-5 Beta Test

by Thomas DeBellis

Version 5.3 is the first major release of Kermit-20 in two decades and represents over a year of development for the following functionality:** ·Updates Kermit to version 7 of Tops-20.** ·Adds DECnet NRT transport to Tops-10, Tops-20 and Ultrix hosts.** ·Adds pseudo-terminal support.** ·Support for batch stream execution, including extensive testing.** ·Microsecond timing (limited by Tops-20)** ·Efficiency Enhancements** ·Symbolic (C) escape sequences and enhanced parsing** ·New and Enhanced commands** ·Revamped and enhanced macro facility** I am still working out where this is going up on the Internet, but for now, HECnet users can feel free to access the package anonymously as follows: VENTI2::K20: Kermit-20 Executable and release notes VENTI2::K20D: Kermit documentation in various formats VENTI2::K20S: Macro source code plus build files for Batch (.CTL) and Interactive (.MIC) VENTI2::K20T: Extensive testing control files and results *Gigantic* thanks to Bob Armstrong for providing /countless/ Tops-10 cycles, support and other resources to me for testing and troubleshooting. I'm celebrating climbing out of this particular rabbit hole by breaking Tops-20 NFT, but FAL is OK. I think.

2 years

Searching for language/compiler

by Johnny Billquist

I wonder if anyone can help me remember/locate a language/compiler for RSX that some third party did. It was allowed to be used on a hobbyist basis in the early 2000s, if I remember right. My brain says XDT or GCML as the company, but searching for that gave me nothing. The language might have been something like RPG II, but again, I am not at all sure my brain is remembering things right. I thought I also had it downloaded and stashed away somewhere, but cannot find anything now that I'm searching. So now I'm hoping this things a bell for someone, somewhere, and I can get my brain sorted, and another tool added on Mim (eventually). :-) Johnny -- Johnny Billquist || "I'm on a bus || on a psychedelic trip email: bqt(a)softjar.se || Reading murder books pdp is alive! || tryin' to stay hip" - B. Idol

2 years

RSX-11 SYSGEN (NOOB) QUESTION

by Mark Curtis

Hi All, I've dusted off my PiDP11/70 running RSX-11M+ (v4.6). Anyway the "stock" image that came from the PiDP site is not to my taste so I've been trying to customise it. Nothing fancy just changing the system name from the standard "PiDP11" to "WITHER". I should stress that I am an utter noob and have zero experience with RSX-11 so doubtless I will use the wrong terms and I know my issues are due to that lack of understanding. To make my life easier with SYSGEN I was using the existing answer files (SYSGENSA1.CMD, etc) and simply editing the items I want. The trouble is, generating a system with these files is producing a system image that has a very small secondary pool (SECPOL). That then produces various errors with commands grumping that there is insufficient SECPOL.... I have figured out how to expand the secondary pool from the command line (which prevents any grumping) but I would like to permanently fix the situation. From the documentation that I've read, and as I understand it, I should be able to pause the SYSGEN and edit files manually mid-generation. In this case I should be able to edit the "SYSVMR.CMD" which is generated at an earlier phase within SYSGEN and then continue the SYSGEN which will then utilise the file (SYSVMR.CMD) in one of the later phases . Except, it does not seem to work. Editing the SYSVMR.CMD file prior to starting SYSGEN clearly overwrites the changes during an early stage. Pausing the SYSGEN and editing the SYSVMR.CMD file mid-generation does retain the settings but does not appear to make any difference to the final image. By way of a little more explanation. I'm presuming once I've "paused" the SYSGEN all I need to do is EDT the SYSVMR.CMD file (in other words, I'm not doing anything further, running any other commands or what-not). All I'm doing is editing the SYSVMR.CMD file (specifically the "SET /PAR=SECPOL" line), and restarting the SYSGEN with "UNS AT.<ESC>". I presuming I am missing out either a vital step or command or completely misunderstanding the "set /par=secpol" line. Any pointers gratefully received Mark

2 years

KL10/RP06/TU77 Power requirements

by Thomas DeBellis

I got asked what the power requirements would have been for 1985 vintage KL10 based system. Does anybody know where I would be able to find such information? I had thought it was in the site installation guide, but I haven't found an online version, yet. About the only thing I can remember is that it was three phase power and that the power supplies were very inefficient.

2 years

Bard writes a song about DECnet

by Supratim Sanyal

I asked Bard to “Write a song about digital equipment corporation’s DECnet network”. The AI is surprising, the ending surreal. (Verse 1) Digital Equipment Corporation, DECnet was your creation. A suite of network protocols, To connect your computers. (Chorus) DECnet, DECnet, It was the best network around. DECnet, DECnet, It connected us all. (Verse 2) You could connect your PDP-11s, To your VAX/VMS machines. You could share files and printers, And communicate with ease. (Chorus) DECnet, DECnet, It was the best network around. DECnet, DECnet, It connected us all. (Bridge) But then the internet came along, And DECnet fell out of favor. But it will always be remembered, As the network that connected us all. (Chorus) DECnet, DECnet, It was the best network around. DECnet, DECnet, It connected us all.

2 years

Copying known nodes command.

by Brian Roth

Hello all, Just added a simh vax 11/780 to HECnet node 31.270 (FATTY) . Circuit is up. Using the copy known nodes command I get "unable to connect to listener" error. Any thoughts on this? Thank you, Brian.

2 years, 1 month

Re: DECnet/8 (was: Re: Copying known nodes command.) --> Mail to Paul @18-nov-2021> RE: Decnet8 and Pydecnet met Pdp8 --> update 8

by R. Voorhorst

L.S. ( & Paul), I sent the mail below (in Dutch) to Paul regarding Decnet-8. In short the conclusions translated: 1. I ironed out some minor bugs and got it running on a Simh Pdp8A simulator with KT8 in Km8 mode with 32 kW memory on Rts8-V2b and with Os8 in 12 kW in background. It has working basic Tlk/Lsn and Nip functionality (see display of Nip below) based on serial lines Ttix/Ttox and with/without Kg8 crc support. It is not very useful beyond that, unless provided with customer programming, and the ddcmp communication can easily be disrupted/poisoned by introducing incomprehensible net traffic packages from Pydecnet (routing messages); the error recovery is not well enough or absent and will kill the line communications leading to a software reboot. Within the Pdp8 framework it is stable (enough?) though. Decnet8 should also work with Dp01/Dp8E sync lines but these are not (yet) implemented and with the DKC8-A parallel interface (also not implemented) 2. This config could be reconfigured and made runnable on a standard Simh Pdp8. As only Decnet8 packet exchange is looked for, the original setup will do. 3. For future expansion the standard max 32 kW memory will not be enough and for Pdp8a Kt8 with up to 128 kW support, Decnet-8 has to be converted to Macrel/Link code within Rts8-V3. It could run with Os8 background in 342 kW and with a bank-contained Fpp8 in another 32 kW it leaves 65 kW for Rts8 and Decnet8. That conversion is not difficult to do but a job which has to be (and will be) done soon enough. 4. To make it more usable, Decnet8 will have to be lifted to Phase-II which is somewhat :( more work. Not too much differences in Ddcmp but much more in Nsp It is all a matter of available resources. 5. I can provide Paul with the basic config that still runs over here so that he can play around. Best regards, Reindert -----Original Message----- From: Robert Armstrong [mailto:bob@jfcl.com] Sent: Monday, 13 March, 2023 20:12 To: 'The Hobbyist DECnet mailing list' <hecnet(a)lists.dfupdate.se> Subject: [HECnet] Re: DECnet/8 (was: Re: Copying known nodes command.) >Paul Koning <paulkoning(a)comcast.net> wrote: >I just went back to the "decnet8.doc" file which is a DECnet for RTS-8 >user manual and internals document. Among other things, it describes >the protocol (DDCMP and NSP). > >I don't know much about PDP-8 software, but if someone manages to get >that code running I'll take a stab at PyDECnet support for talking to it. Can pyDECnet do asynchronous DECnet over a telnet port? I'm thinking the easiest way to set this up for debugging/testing would be to use simh for the PDP-8 and a virtual KL8-E connected to pyDECnet. Ultimately it would be nice to run it on real hardware with a real async DDCMP connection, but that's for later. Bob _______________________________________________ HECnet mailing list -- hecnet(a)lists.dfupdate.se To unsubscribe send an email to hecnet-leave(a)lists.dfupdate.se **************************************************************************** ******************************** Dag Paul, Er is inmiddels weer wat water door de rivieren gestroomd en ik heb basis decnet8 werkend gekregen: ****************** NETWORK INFORMATION PROGRAM V1D PHYSICAL LINE STATUS [DDCMP V1A] LINE NODNAM (#) STATE DRIVER VRSN UP SINCE 0 SWBZ04 (161) ON-LINE KL8J 1A 0:46:13 17-NOV-76 1 SWBZ02 (157) OFF-LINE KL8J 1A LOGICAL LINK STATUS [NSP V1C] CHAN STATE LINE TASK SRCE-NAME[PPN] TYPE LINK DEST-NAME[PPN] TYPE LINK 1 UNUSED 0 35 TLK [1,3] 0 6 2 UNUSED 0 36 TLK [1,2] 0 2 PHYSICAL LINE ERROR STATUS LINE 0: NO ERRORS 16 MESSAGES TRANSMITTED 12 MESSAGES RECEIVED LINE 1: NO ERRORS 0 MESSAGES TRANSMITTED 0 MESSAGES RECEIVED ******************* Het bleek nog niet eens zo moeilijk te zijn meer een zaak hoe de decnet configuratie volgens de lokale spelregels met eigenaardigheden om te zetten in assembler parameters en dan het rts-8 systeem te bouwen. De onderlaag lijkt (voorlopig) goed genoeg te werken, maar de cliënten TLK/LSN voldoen niet compleet aan hun specs. Maar ook de selectie van node naam blijkt niet afdoende; een pakket naar een node die niet adjacent is wordt gewoon door de adjacent node als bestemming geaccepteerd terwijl de bestemming node naam niet de juist is. Met name het zenden van berichten naar de eigen node werkt niet en dus ook de default commando's niet waar de eigen node naam wordt geëvalueerd. Het lijkt dat decnet8 de mogelijkheid van een alias definitie niet ondersteunt. Ook lijkt 0:: niet een optie te zijn als alias maar zou misschien wel moeten. Wat betreft de verdere weg is het duidelijk dat Decnet8 alleen zal werken binnen de niche Pdp8 en dat communicatie met andere types decnet nodes uitgesloten is. Dit betekent m.i. dat voor afdoende Decnet8 functionaliteit het product naar Phase-II moet worden opgetild met eventueel verzwakte functionaliteit door het weglaten van niet benodigde deel functies (Ddcmp MOP boot?). Dat betekent dat een Phase-II compatibele Decnet8 kan communiceren met de andere Phase-II nodes zoals al in aanleg aanwezig moet zijn geweest vlak voordat het product afgedankt werd. Zouden er nog ergens kopieën van rondzwerven? Enig idee? In elk geval betekent dit geen wijzigingen voor Pydecnet aangezien ik aanneem dat er verder geen Phase-I/? ondersteuningswensen voor Pydecnet aanwezig zullen zijn? Ik ga ervan uit dat Pydecnet wel de Phase-II intercept ondersteunt, anders moet er een Dec variant bestaan die dat kan invullen. Zijn er nog wetenwaardigheden die in acht moeten worden genomen voor het (eventuele) omzetten naar Phase-II? Decnet8 intercept zal m.i. niet tot het aangeboden palet gaan behoren. Kan Pydecnet omgaan met async lijnen die naar Tcp/ip sockets i.p.v. com poorten afgebeeld zijn? Nu lijkt het erop dat alleen synchrone lijnverbindingen werken in die modus wat een beperking inhoudt. Reindert **************************************************************************** ****************************

2 years, 1 month

← Newer
1
...
7
8
9
10
11
12
13
...
1556
Older →

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

HECnet