[Info-vax] SET TERM /TTSYNC on ancient VMS versions

[George Cornelius]

In article <pl4spa$1r0v$1 at gioia.aioe.org>, Chris <xxx.syseng.yyy at gfsys.co.uk> writes:

> Common in industry, which still uses a lot of serial line
> hardware. A common pair being rts / cts, to limit data rates from
> host to a peripheral. RS232 only over short distances, the 423 / 485
> differential for greater distances.
> 
> The early uart devices, also used by dec, had no buffering, so
> as soon as a char arrived, you got an interrupt or had to use a polling
> loop. Later devices had multibyte buffering, so you had a bit more
> time to read the byte and decide what to do with it, important at the
> higher baud rates. The later dec serial cards used custom
> microcircuits and no idea what they had on chip, but probably used dma
> by then...

We are mostly running in circles. And it's close to being moot
given that real serial ports are going the way of the dodo.

I believe the I/O Users' Guide (is that what it was called?)
clearly specified that hardware flow control was not supported.

But my recollection is that, for the most part, it worked.

I strongly suspect, by the way, that characters like
CTRL/C and CTRL/Y are treated as out of band, in which
case they would with software flow control be able to
affect program execution even if normal character flow
were blocked.  Not so with hardware flow control, which
is not capable of making such distinctions.

Otherwise hardware flow control, for properly wired
cables, would appear to be superior.

George
 
> Chris
>