13 Mar
2023
13 Mar
'23
2:44 p.m.
On Mar 11, 2023, at 9:33 PM, Thomas DeBellis
<tommytimesharing(a)gmail.com> wrote:
It would be interesting to confirm that. If it's true, then there is a potential
answer: implement Phase IV Prime. That's a thingly documented extension of Phase IV
originally done to support IEEE 802.5 token ring (with its inaccurately named
"functional addresses"). Among other things it supports using the
"hardware address" rather than the Bernie Lacroute "hiord" address.
I'm planning to implement that at some point in PyDECnet to allow it to talk to DECnet
nodes on WiFi, where changing the address doesn't normally work. But if there is any
other hardware with that limitation it would work there as well. The changes are pretty
modest, especially for end nodes.
The protocol details are in the 802.5 datalink architecture spec (EK-DNAP4-TR-001), of all
places. I forgot where I found that document online, the filename is
"trn_dlspec_v100.pdf".
paul
On 3/8/23 5:13 PM, Keith Halewood wrote:
HECnet is a wonderful achievement. I'm not the only one who tried, but it would have
been great to get LCM+L's DEC systems connected, Rosie in particular. I have also
cajoled the 'twenex' people and also those in (UK) National Museum of Computing
without success.
...
Apparently not all 36 bit hardware will do DECnet. I can't remember exactly, but I
believe the network interfaces on some Toad's can not have their MAC addresses
programmatically changed (anybody know this?)