Concerned Mike writes:
 
[snip]
 
>
> We all know that when the HP3000 is sending stuff too fast for us to
> read, we can do ^S to stop the data, and ^Q to resume it. This shows
> that the HP3000 does software flow control, using XON (^Q) and XOFF
> (^S). However, this only proves that the HP3000 provides this flow
> control in the transmit direction (from the HP3000 to your terminal),
> listening for and taking action on your XON/XOFF commands. The weird
> part is that it appears to me that the HP3000 does not use any flow
> control at all in the receive direction (from your terminal to the
> HP3000).
>
> I configured my modem to act on XON/XOFF characters sent by the
> HP3000 to the modem, on the theory that these fast modems might send
> data to the HP3000 faster than it can digest, and that the HP3000
> would signal the modem to slow down by sending an XOFF. This caused
> no end of trouble, particularly with the terminal status checks that
> Qedit does. The usual result was that the terminal keyboard stayed
> locked, and Qedit would assume that it was working with a non-HP dumb
> terminal. I finally setup the modem to send XON/XOFF characters to
> the HP3000 as required, but not to listen for them from the HP3000.
> This seems to allow most everything to work as desired.
>
> I am left with the nagging worry that as these faster modems (> 19200
> baud) become more common, we will eventually run into cases where the
> HP3000 cannot digest the data quickly enough, but cannot signal the
> modem to pause and resume.
>
> Is my concern justified, considering that the baud rate between the
> modem and the HP3000 is only at 19200, even if the two modems
> communicate at a faster rate?
>
> Regards, Mike Shumko, Robelle Consulting, 1-800-561-8311
>
 
Rest easy Mike! Let's consider the source of the data flowing from your
PC to the 3000: 1 - your fingers, 2 - the enter key, 3 - a file transfer.
1) Ha! You wish you could type so fast!
2) OK, your mega VPlus screen sends 2000 characters, could that be a
problem? I suppose it could, though I've only ever been able to get the
DTC to generate an XOFF when the type ahead buffer fills up. If a read
is posted, the DTC unloads data to the 3000 at about 250 characters a pop
and has about 1000 characters of buffer.
3) Essentially the same as 2, the file transfer programs do not send a
continuous stream of data, they send a block, get an ack, repeat.
 
If you're really concerned, tell the modem at the 3000 end to obey
XON/XOFF and pass them on to your PC. Your problems were caused by the
modem eating the DC1s (XONs) that trigger terminal reads. In XON/XOFF
pass-through, the modem will obey oodles of XONS and pass them on to
keep your PC happy. Once in a great while, it might actually see an XOFF.
 
--
--------------------------------------------------------------------
Ross Scroggs                         email: [log in to unmask]
Telamon, Inc.                          CIS: 76011,2234
492 Ninth Street, Suite 310          voice: 510-987-7700
Oakland, CA 94607-4098                 fax: 510-987-7009
--------------------------------------------------------------------