[ih] NCP and TCP implementations

[Jack Haverty]

We had a lot of trouble with LSI-11 devices inserted between a host and
an IMP.   Your experience may have been similar.   I even wrote a note
about the problem:

https://www.rfc-editor.org/ien/ien139.txt

The 1822 spec was about how to interface a Host to an IMP.  People
assumed the spec was symmetrical, but it wasn't.  An IMP had to do more
than a Host.   So you couldn't quite just plug two "1822" interfaces
together and get it to work reliably.

/Jack


On 7/22/20 8:35 AM, Noel Chiappa via Internet-history wrote:
>  I did try and get the Port Expander working,
> but because someone who shall remain nameless, but I'm sure knows who he is,
> had done an analog hack on DM's ARPANet interface [something to do with
> grounding on the DH differential interface, IIRC], our LSI-11 ARPANET
> interfaces (the SRI design, that used a DRV11), which didn't include the
> opto-isolators of the IMP DH interface, wouldn't talk to the DM one, so that
> was the end of that!