On Monday, September 14, 2020 at 2:24:59 PM UTC-7, AB wrote:
> Hello everyone- I seem to have hit a hard limit to the number of USB-connected terminals using TOPS-10-in-box, under the latest SIMH on a Pi 4.
> Seems to max out after 8 devices (ttyUSB0-7) and I was wondering if that's a configurable maximum in SIMH... or something else?

This is reminding me of TOPS-10 stories from about 50 years ago.

It seems that TOPS-10 gets an interrupt for every character typed, and that if you type fast enough, possible to overflow some buffer.
The story I heard was that there was someone who could type that fast. (I believe a Diablo terminal at 300 baud.)

But already that long ago, and more common now, is to connect a computer to a serial port, where it can run at full speed.
You ask about the number of ports, but not the data rate for those ports.

gah4 <gah4@u.washington.edu> writes:

[snip]

> This is reminding me of TOPS-10 stories from about 50 years ago.
>
> It seems that TOPS-10 gets an interrupt for every character typed, and
> that if you type fast enough, possible to overflow some buffer.
> The story I heard was that there was someone who could type that fast.
> (I believe a Diablo terminal at 300 baud.)

This might be a garbled version of a story that Greg Zima (of Home
Office Software Support) started. He posted an April 1st report of
a site visit to the Metropolis Daily Planet, where an anonymous fast
typist was able to generate BIGBFO BUGCHKs on a TOPS-20 system.
(BIGBFO = BIGBUF overflow). The terminal in question would have been
the VT-52, which had been observed to be present in the new (1978)
Superman movie.

Scott
--
Scott Hemphill hemphill@alumni.caltech.edu
"This isn't flying. This is falling, with style." -- Buzz Lightyear

On Tuesday, April 25, 2023 at 11:22:06 AM UTC-7, Scott Hemphill wrote:

(I wrote)

> > It seems that TOPS-10 gets an interrupt for every character typed, and
> > that if you type fast enough, possible to overflow some buffer.
> > The story I heard was that there was someone who could type that fast.
> > (I believe a Diablo terminal at 300 baud.)

> This might be a garbled version of a story that Greg Zima (of Home
> Office Software Support) started. He posted an April 1st report of
> a site visit to the Metropolis Daily Planet, where an anonymous fast
> typist was able to generate BIGBFO BUGCHKs on a TOPS-20 system.
> (BIGBFO = BIGBUF overflow). The terminal in question would have been
> the VT-52, which had been observed to be present in the new (1978)
> Superman movie.

In my case, I actually knew the person, though didn't personally
see it happen. It was on a KA-10 and about 1975.

On Tuesday, April 25, 2023 at 11:22:06 AM UTC-7, Scott Hemphill wrote:

(snip)

> Scott Hemphill hemp...@alumni.caltech.edu
I just realize that you were closer to the event than I was.

Maybe you remember a room with mostly DECwriters, one Diablo,
and one Tektronix 4010. I wasn't there until 1976, though, and it was
supposed to be a little earlier.

On Monday, April 24, 2023 at 11:31:39 PM UTC-4, gah4 wrote:
> It seems that TOPS-10 gets an interrupt for every character typed, and that if you type fast enough, possible to overflow some buffer.
> The story I heard was that there was someone who could type that fast. (I believe a Diablo terminal at 300 baud.)

Yes, both TOPS-10 and Tops-20 could run into this. Tops-20 at least had the TRMOP bits that could be set to
let the front-end PDP-11 just echo simple characters. But teco and later tv, would change the action of printable
characters based on whatever mode you were in.

I don't think human typists could actually trigger it. (I did have a systems programmer who was trained as a classical pianist, he could really press the keys, he could type 200 words per minute, which was very fast). But in the last 70s, a lot of vendors
moved from simple glass teletypes like the VT52, VT100, ADM-3, etc. and supported a "forms" based display
with prompts and blanks for entering business forms. Most of these just buffered up the input, and submitted it all at 1200 or 2400 bps when you hit enter. This could over run the processing unless the form submission supported X-on/X-off (^S ^Q)

Other systems, like the HP3000 used a more sane protocol, the "enter" sent a string to the host, which would allocate a 2000 byte buffer, and then tell the terminal to dump its local buffer down the serial line.

On Wednesday, April 26, 2023 at 3:24:09 PM UTC-7, fishtoprecords wrote:
> On Monday, April 24, 2023 at 11:31:39 PM UTC-4, gah4 wrote:
> > It seems that TOPS-10 gets an interrupt for every character typed, and that if you type fast enough, possible to overflow some buffer.
> > The story I heard was that there was someone who could type that fast. (I believe a Diablo terminal at 300 baud.)

> Yes, both TOPS-10 and Tops-20 could run into this. Tops-20 at least had the TRMOP bits that could be set to
> let the front-end PDP-11 just echo simple characters. But teco and later tv, would change the action of printable
> characters based on whatever mode you were in.
> I don't think human typists could actually trigger it.

The one I knew was on a KA-10, which might matter.

(snip)
> Other systems, like the HP3000 used a more sane protocol, the "enter" sent
> a string to the host, which would allocate a 2000 byte buffer, and then tell the
> terminal to dump its local buffer down the serial line.

For serial lines, IBM S/360 and S/370 only interrupt every line.

The IBM favorite, which I mostly didn't use, is the 3270 terminal, where the terminal
buffers the whole screen, and then transfers whatever is needed as one operation.