11 Dec
2019
11 Dec
'19
4:03 a.m.
I would say it because of the code snippet at the bottom of G.'s,
letter, viz:
FAL$RENAME:: ; Entry point BBS #DAP$V_GEQ_V70,(R9),10$ ; *Exit if
partner does not support* BRW FAL$UNS_ACCFUNC ; *DAP V7.0*
The first instruction is a branch-on-bit-set and it looks like it is
testing a bit and branching if it is set. It would appear that bit is
associated with V7 DAP, based both on the symbolic value and comment. If
that branch isn't taken, then an unconditional branch is taken to what
appears to be an unsupported function return. However, I am in no way a
VAX assembly hacker. We looked it briefly when our chemistry department
got their first VAX and decided to stay with high level languages. While
the VAX had some fun stuff, the PDP-10's instruction set was considered
far easier; definitely the paging subsystem. However, I suspect that was
more because we had decades of experience with it. The VAX multiple byte
architecture had some remarkable similarities to a number of 360, 3033
and 4341's that we had, which weren't as much fun to program in
assembler. At any rate, that's why I said, "my guess".
On 12/10/19 8:02 PM, Paul Koning wrote:
On Dec 10, 2019, at 6:36 PM, Thomas DeBellis
<tommytimesharing at gmail.com> wrote:
Interesting.
Do you have source for the DAP/NFT/FAL on Tops-10 or know where I might get what you
have?
May I ask you what it does against a VMS system? Unless it is reporting DAP 7, my guess
is that it won't work.
Why would you say that? The DECnet rule is backward
compatibility, that's why the version numbers are exchanged.
Actually, DAP takes this very seriously indeed, which is why it sends very long bitmaps
describing in detail all the protocol features supported.
It may well be that a V7 DAP implementation won't talk to V4 (Phase II). But I'd
be very surprised to see it fail talking to a V5.6 protocol peer -- I'm pretty sure
RSTS uses that version and we would certainly have complained if other Phase IV
implementations had trouble with that.
More in general, across the board the rule always was "One version back
compatible". So Phase IV can talk to Phase III (but not to Phase II -- except for
PyDecnet. :-) )
paul