[HECnet] Disallow login on DZ line in VMS

[Thomas DeBellis]

Have you ever gotten the impression that terminal support sometimes just 
'happens' instead of being planned?  The amount of parameters and 
options that you can set on a line for Tops-10, Tops-20 and Unix, often 
with a number of ways of doing what appears to be the same thing, is 
truly bewildering.

Aside from the names to do the job, I don't think VMS is any worse or 
much better than anything else.  There are plenty of parameters in 
Tops-10 that require about the same amount of explanation (TTY SLAVE and 
TTY GAG come to mind).

For Tops-20, it is less intuitive; the only way to do this is at system 
start up.  Otherwise there are no interactive surfaces to do the 
configuration, meaning SETSPD has no offset in its entry vector for 
parsing from the terminal (which I find remarkable).  I /think/// the 
following line in SYSTEM:7-1-CONFIG.CMD would suffice:

    TERMINAL /XX/ [/- YY/] IGNORE-SYSTEM-MESSAGES IGNORE-USER-MESSAGES
    NOBELL SPEED /zz/

Notes:

 1. NOBELL is key to prevent some kinds of gubbish from interfering with
    the application or protocol
 2. Ditto ignoring system and user blat
 3. The terminal should still be assigned as a _^C_ will cause a job
    creation
 4. REMOTE is probably not indicated unless the device handles DSR (or
    carrier)
 5. SPEED can be split between input and output, but is limited to 9,600
    baud on a DH or DZ

The application doing the device assign could set all of these itself, 
but would need capabilities to do some of them (I forget which).  You 
could then use NCP to set the rest of the LINE parameters, hopefully.  
Since Tops-20 NCP can't assign a line, what you could do is have the job 
that starts the NCP issue an ASSIGN command beforehand (/after/ SETSPD 
and /before/ any logins are allowed).

Kind of baroque, I'd have to say.  I'd also say that VMS, RSX and VMS 
user interface is probably more straightforward, in this regard (absent 
the goofy naming)

Tops-20 appears like it might have had some hooks doing these kinds of 
things for application terminals, but I don't see any vestigial or 
otherwise unfinished code for the conditional, so I don't know what they 
might have done.  Intriguing...

On 12/21/21 9:14 AM, Johnny Billquist wrote:
> Like I said - I do find it a bit unintuitive that all of this 
> processing is so tied to the type ahead attribute. 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.sonic.net/mailman/private/hecnet-list/attachments/20211221/f4705323/attachment.htm>