On 2022-06-16, Charles Richmond <codescott@aquaporin4.com> wrote:

> At a PPOE, I used a Digital Equipment VT-220 which was attached via a
> termserver to an ethernet network to rlogin to one of the workstations.
> This terminal could keep up with listing out a file at 19200 baud...
> as long as all the lines had only 40 characters or less. If the file
> had a grouping of lines with 70 or 80 characters, the speed of the
> display would bog down...

Interesting. I was thinking of overhead involved in scrolling the
screen, but this sounds different. Maybe it was smart enough to
recognize short lines and not bother shifting blank areas.

The Univac mainframe terminals I worked with had serious scrolling
overhead; they would take 20 milliseconds to scroll an entire screen
of data, and they were blind to incoming data during that time.
At 9600 bps you had to pad each scroll sequence with 20 NULs
to avoid data loss.

> I did *not* like the 220 for many reasons... not the least being that
> the escape key was badly placed. We were running SVR4 Unix. But the
> terminal had a monochrome amber display.

The escape key was the least of your worries. Although the VT-100
had quite a nice keyboard, the designers dropped the ball on the
VT-220 and successors, incorporating the worst misfeatures of the
original IBM Personal Computer keyboard: the extra key between Z
and shift, and the remotely-placed return key.

Fortunately I never had to use any of the later DEC terminals.
The horrors that Univac perpetrated on their later keyboards,
however, were even worse.

IBM has a lot to answer for: in this case, setting keyboard
design back 10 years.

--
/~\ Charlie Gibbs | Microsoft is a dictatorship.
\ / <cgibbs@kltpzyxm.invalid> | Apple is a cult.
X I'm really at ac.dekanfrus | Linux is anarchy.
/ \ if you read it the right way. | Pick your poison.

John Goerzen <jgoerzen@complete.org> wrote:
> On 2022-06-15, Jan van den Broek <fortytwo@xs4all.nl> wrote:
>>> Like I said, I'm not looking to buy anything yet but if you have any
>>> suggestions on what to keep an eye out for, please let me know.
>>
>> Not a dumb terminal, I'm using a 386sx (MS-DOS 5.0) with Telix.
>
> How's its performance?
>
> I've been somewhat astonished at how horribly slow the real dumb terminals are.
>
> Even my vt510, which seems to be the fastest, has to use XOFF when I start emacs
> because it can't keep up with the escape sequences at that point and drops them,
> unless I set it all the way down to 9600bps (it will support line rates up to
> 115200bps). Something of an additional annoyance is that these real dumb
> terminals support hardware flow control, but using the DTR pin, not RTS/CTS, and
> that isn't supported on Linux (though it is on FreeBSD).
>
> My Wyse WY-55 is worse than the vt510, and my IBM 3151 won't even support a line
> rate above 19200.
>
> John
>

The escape sequences were a bit slower than just displaying data, but what
could you DO with a faster terminal? How fast can you read data as it
scrolls by on the screen?

--
Pete

I wrote a terminal emulator that ran on Datapoint workstations once...
it turned out there was an array of pointers, one for each line of
text to display (graphics? Surely you jest. This was a business
machine). Things like scrolling were much quicker than they'd have been
if I'd actually had to move the text around.

On Wed, 15 Jun 2022 20:55:19 -0000 (UTC)
John Goerzen <jgoerzen@complete.org> wrote:

> Something of an additional annoyance is that these real dumb
> terminals support hardware flow control, but using the DTR pin, not
> RTS/CTS, and that isn't supported on Linux (though it is on FreeBSD).

There's a reason so many of us carried RS232 break out boxes, patch
boxes, gender benders etc. Black Box did a very handy set back in the day,
I saw some bits of mine a couple of years ago.

--
Steve O'Hara-Smith
Odds and Ends at http://www.sohara.org/

Ahem A Rivet's Shot <steveo@eircom.net> writes:

> On Wed, 15 Jun 2022 20:55:19 -0000 (UTC)
> John Goerzen <jgoerzen@complete.org> wrote:
>
>> Something of an additional annoyance is that these real dumb
>> terminals support hardware flow control, but using the DTR pin, not
>> RTS/CTS, and that isn't supported on Linux (though it is on FreeBSD).
>
> There's a reason so many of us carried RS232 break out boxes, patch
> boxes, gender benders etc. Black Box did a very handy set back in the day,
> I saw some bits of mine a couple of years ago.

I think I still have mine... hmmm.... yep. Just found it. Made by
Inmac.

On Thu, 16 Jun 2022 18:30:11 -0600
Joe Pfeiffer <pfeiffer@cs.nmsu.edu> wrote:

> Ahem A Rivet's Shot <steveo@eircom.net> writes:
>
> > On Wed, 15 Jun 2022 20:55:19 -0000 (UTC)
> > John Goerzen <jgoerzen@complete.org> wrote:
> >
> >> Something of an additional annoyance is that these real dumb
> >> terminals support hardware flow control, but using the DTR pin, not
> >> RTS/CTS, and that isn't supported on Linux (though it is on FreeBSD).
> >
> > There's a reason so many of us carried RS232 break out boxes,
> > patch boxes, gender benders etc. Black Box did a very handy set back in
> > the day, I saw some bits of mine a couple of years ago.
>
> I think I still have mine... hmmm.... yep. Just found it. Made by
> Inmac.

Oh ye glods Inmac - for a while I got three of their catalogues at
a time because they hated to lose any circulation.

--
Steve O'Hara-Smith
Odds and Ends at http://www.sohara.org/

Charlie Gibbs <cgibbs@kltpzyxm.invalid>, 2022-06-16 05:00 (+0000):

> Although the VT-100 had quite a nice keyboard, the designers dropped
> the ball on the VT-220 and successors, incorporating the worst
> misfeatures of the original IBM Personal Computer keyboard: the extra
> key between Z and shift, and the remotely-placed return key.

There was the DEC LK421 "Unix keyboard", though:

https://en.wikipedia.org/wiki/File:LK421.jpg

--
MC, https://hack.org/mc/

Joe Pfeiffer wrote:

> I wrote a terminal emulator that ran on Datapoint workstations once...
> it turned out there was an array of pointers, one for each line of
> text to display (graphics? Surely you jest. This was a business
> machine). Things like scrolling were much quicker than they'd have been
> if I'd actually had to move the text around.

Was that on a 2200/5500/6600 series, or a 1500, or an 1800/3800 series.
Maybe I forgot some of them.

--
Carefully cut along this line.

/|>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>\|
/| I may be demented \|
/| but I'm not crazy! \|
/|<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<\|
* SPAyM trap: there is no X in my address *

Freddy1X <freddy1X@indyX.netX> writes:

> Joe Pfeiffer wrote:
>
>> I wrote a terminal emulator that ran on Datapoint workstations once...
>> it turned out there was an array of pointers, one for each line of
>> text to display (graphics? Surely you jest. This was a business
>> machine). Things like scrolling were much quicker than they'd have been
>> if I'd actually had to move the text around.
>
> Was that on a 2200/5500/6600 series, or a 1500, or an 1800/3800 series.
> Maybe I forgot some of them.

I'm thinking it was on a 2200 series, but I couldn't swear to it. It
was 1980....

Joe Pfeiffer wrote:

> Freddy1X <freddy1X@indyX.netX> writes:
>
>> Joe Pfeiffer wrote:
>>
>>> I wrote a terminal emulator that ran on Datapoint workstations once...
>>> it turned out there was an array of pointers, one for each line of
>>> text to display (graphics? Surely you jest. This was a business
>>> machine). Things like scrolling were much quicker than they'd have been
>>> if I'd actually had to move the text around.
>>
>> Was that on a 2200/5500/6600 series, or a 1500, or an 1800/3800 series.
>> Maybe I forgot some of them.
>
> I'm thinking it was on a 2200 series, but I couldn't swear to it. It
> was 1980....

In 1980 I believe that the 2200 & 1100 were the only processors that
Datapoint had out at the time. Your view was 12 rows of 80 characters each.
Invariably green phosphor. ISTR that the 1100s were paired up with 8"
diskette drives, ant the 2200 used dual cassette decks.

Was the display board the serial or RAM version?
The serial display board had a bizarre way of updating the displayed
characters depending on how the program ran it. And it was SLOW. So much
so that when I had to run diagnostics on the tape decks, I would swap in the
RAM display board just so that I could be done in a reasonable time.

I never got to play with them for any length of time because I was a 1 man
remote office, driving to location as repair service was needed.

Freddy,
Datapoint field circus, 1979 to 1990.
--
You are not a winner if your game piece reveals "PLEASE TRY AGAIN".

/|>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>\|
/| I may be demented \|
/| but I'm not crazy! \|
/|<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<\|
* SPAyM trap: there is no X in my address *

Wed, 18 May 2022 11:33:08 -0400
Andreas Kohlbach <ank@spamfence.net> schrieb:
>On Wed, 18 May 2022 00:09:54 -0000 (UTC), Chris Adams wrote:
>>
>> Note that the Pi serial port is not going to work directly on a classic
>> terminal. The Pi signalling is not RS-232 (typically +12/-12V IIRC) but
>> 3.3V/0V. Sending more than 3.3V into a Pi serial pin will fry it.
>>
>> There are pre-made Pi-to-USB serial adapters for connecting to a PC, but
>> I don't know if there are pre-made Pi-to-read-RS-232 adapters.
>
>Curious, somebody threw out a 2011 LED TV. One jack on the backside is
>labeled RS-232. What would you do with that on a TV?

Turning it on and off, switching input, changing channels/volume/etc.
At a ppoe we were maintaing a narrowcasting application, rs232 was used
at lot.

>It also had some USB jacks; that I can understand.
--
A tuna is a way of Liff

Jan v/d Broek
balglaas@xs4all.nl

Wed, 15 Jun 2022 20:55:19 -0000 (UTC)
John Goerzen <jgoerzen@complete.org> schrieb:
>On 2022-06-15, Jan van den Broek <fortytwo@xs4all.nl> wrote:
>>>Like I said, I'm not looking to buy anything yet but if you have any
>>>suggestions on what to keep an eye out for, please let me know.
>>
>> Not a dumb terminal, I'm using a 386sx (MS-DOS 5.0) with Telix.
>
>How's its performance?

Depending on what you want. I'm using it at 19200, which is fast enough
for me. I mainly use it for editing, compiling, usenet.

[Schnipp]
--
A tuna is a way of Liff

Jan v/d Broek
balglaas@xs4all.nl

On 2022-06-21, Jan van den Broek <fortytwo@xs4all.nl> wrote:
> Wed, 15 Jun 2022 20:55:19 -0000 (UTC)
> John Goerzen <jgoerzen@complete.org> schrieb:
>>On 2022-06-15, Jan van den Broek <fortytwo@xs4all.nl> wrote:
>>>>Like I said, I'm not looking to buy anything yet but if you have any
>>>>suggestions on what to keep an eye out for, please let me know.
>>>
>>> Not a dumb terminal, I'm using a 386sx (MS-DOS 5.0) with Telix.
>>
>>How's its performance?
>
> Depending on what you want. I'm using it at 19200, which is fast enough
> for me. I mainly use it for editing, compiling, usenet.

I'm trying to remember my days of a 386SX/25 in OS/2. My speeds then were
probably limited by my modem (33.6 at best). I don't recall any obvious PC-side
slowdowns like I see on the vt510.

Also trying to remember the days of a 7.16MHz 8088ish DOS box. I guess I didn't
have a modem faster than 2400bps in those days. Still, I think it was able to
run LapLink at 115200 but I may be somewhat mistaken about that, and that may be
a different problem to ANSI video processing too.

I may have to give this a try one of these days.

John

Freddy1X <freddy1X@indyX.netX> writes:

> Joe Pfeiffer wrote:
>
>> Freddy1X <freddy1X@indyX.netX> writes:
>>
>>> Joe Pfeiffer wrote:
>>>
>>>> I wrote a terminal emulator that ran on Datapoint workstations once...
>>>> it turned out there was an array of pointers, one for each line of
>>>> text to display (graphics? Surely you jest. This was a business
>>>> machine). Things like scrolling were much quicker than they'd have been
>>>> if I'd actually had to move the text around.
>>>
>>> Was that on a 2200/5500/6600 series, or a 1500, or an 1800/3800 series.
>>> Maybe I forgot some of them.
>>
>> I'm thinking it was on a 2200 series, but I couldn't swear to it. It
>> was 1980....
>
> In 1980 I believe that the 2200 & 1100 were the only processors that
> Datapoint had out at the time. Your view was 12 rows of 80 characters each.
> Invariably green phosphor. ISTR that the 1100s were paired up with 8"
> diskette drives, ant the 2200 used dual cassette decks.

They had introduced their ARCNET network at that point (probably the
major reason they were picked), so I don't recall any local disk or
cassette storage at all.

> Was the display board the serial or RAM version?

I have no recollection whatever. I don't even remember that there was a
choice.

On Wed, 22 Jun 2022 23:56:58 -0000 (UTC)
John Goerzen <jgoerzen@complete.org> wrote:

> I'm trying to remember my days of a 386SX/25 in OS/2. My speeds then were
> probably limited by my modem (33.6 at best). I don't recall any obvious
> PC-side slowdowns like I see on the vt510.

PCs tended to come with a buffered UART that cut down the interrupt
rate no end.

--
Steve O'Hara-Smith
Odds and Ends at http://www.sohara.org/

To: John Goerzen
-=> John Goerzen wrote to alt.folklore.computers <=-

JG> I'm trying to remember my days of a 386SX/25 in OS/2. My speeds then
JG> were probably limited by my modem (33.6 at best). I don't recall any
JG> obvious PC-side slowdowns like I see on the vt510.

I ran OS/2 2.1 on a PS/2 model 80 baxk in 1991-92, but I think I only had a
2400 baud modem back then. OS/2 could multitask well, though - I could call
my BBS, connect to a MS Lan Manager network and share my files, connect to a
AS/400 terminal session via twinax, and run Word and Excel natively - with 8
MB of RAM!

JG> Also trying to remember the days of a 7.16MHz 8088ish DOS box. I guess
JG> I didn't have a modem faster than 2400bps in those days. Still, I
JG> think it was able to run LapLink at 115200 but I may be somewhat
JG> mistaken about that, and that may be a different problem to ANSI video
JG> processing too.

It was all about the UART with AT-class machines. With my BBS, I had a
socketed serial card and swapped out an 8250 for a 16550 UART and could run
14.4 all day long. I don't recall if my serial card was 8-bit or 16-bit, and
am unsure if an 8-bit card would be the bottleneck on an XT-class machine.

kurt weiske | kweiske at realitycheckbbs dot org
| http://realitycheckbbs.org
| 1:218/700@fidonet

--- MultiMail/DOS v0.52
--- Synchronet 3.19c-Win32 NewsLink 1.113
* realitycheckBBS - Aptos, CA - telnet://realitycheckbbs.org

To: Ahem A Rivet's Shot
-=> Ahem A Rivet's Shot wrote to alt.folklore.computers <=-

AAS> PCs tended to come with a buffered UART that cut down the
AAS> interrupt rate no end.

Cheap cards typically came with an 8250 or 16450 UART, which had (I think) a
2-byte buffer. You'd need to either spend more to get a card with a 16550
(16-byte buffer) or if you were lucky, upgrade your existing (socketed) card
by swapping out the UART chip.

The 16450 was pretty much useless above 9600 baud.

--- MultiMail/DOS v0.52
--- Synchronet 3.19c-Win32 NewsLink 1.113
* realitycheckBBS - Aptos, CA - telnet://realitycheckbbs.org

"Kurt Weiske" <kurt.weiske@realitycheckbbs.org.remove-11mi-this> writes:
> To: John Goerzen
>-=> John Goerzen wrote to alt.folklore.computers <=-
>
> JG> I'm trying to remember my days of a 386SX/25 in OS/2. My speeds then
> JG> were probably limited by my modem (33.6 at best). I don't recall any
> JG> obvious PC-side slowdowns like I see on the vt510.
>
>I ran OS/2 2.1 on a PS/2 model 80 baxk in 1991-92, but I think I only had a
>2400 baud modem back then. OS/2 could multitask well, though - I could call
>my BBS, connect to a MS Lan Manager network and share my files, connect to a
>AS/400 terminal session via twinax, and run Word and Excel natively - with 8
>MB of RAM!
>
> JG> Also trying to remember the days of a 7.16MHz 8088ish DOS box. I guess
> JG> I didn't have a modem faster than 2400bps in those days. Still, I
> JG> think it was able to run LapLink at 115200 but I may be somewhat
> JG> mistaken about that, and that may be a different problem to ANSI video
> JG> processing too.
>
>It was all about the UART with AT-class machines. With my BBS, I had a
>socketed serial card and swapped out an 8250 for a 16550 UART and could run
>14.4 all day long. I don't recall if my serial card was 8-bit or 16-bit, and
>am unsure if an 8-bit card would be the bottleneck on an XT-class machine.

The 16550 or 16450? The 16550 came later and added a 16 byte FIFO for
inbound characters; the 16450 supported higher data rates than the 8250,
but ultimately was bounded by interrupt service routine performance.

On Fri, 24 Jun 2022 12:52:31 GMT
scott@slp53.sl.home (Scott Lurndal) wrote:

> The 16550 or 16450? The 16550 came later and added a 16 byte FIFO for
> inbound characters;

Then the 16550A came out with the bugs fixed. With them on two 16
port cards I managed to run 32 modems at 19200 on a 50MHz 486 without
dropping a byte.

--
Steve O'Hara-Smith
Odds and Ends at http://www.sohara.org/

Ahem A Rivet's Shot <steveo@eircom.net> writes:

> On Wed, 22 Jun 2022 23:56:58 -0000 (UTC)
> John Goerzen <jgoerzen@complete.org> wrote:
>
>> I'm trying to remember my days of a 386SX/25 in OS/2. My speeds then were
>> probably limited by my modem (33.6 at best). I don't recall any obvious
>> PC-side slowdowns like I see on the vt510.
>
> PCs tended to come with a buffered UART that cut down the interrupt
> rate no end.

Did they? I remember buying an ISA card by SIIG with the 16550 UART that
had the huge 16 byte buffer back in 1993 when I got a 14400 bps
modem. With the V.42bis compression I think you could have at least 2x
compression so double data rate over the serial link. Might've had a 486
by then though.

In my computers the built in ports had just the no buffer
16450s. Same with the multi-I/O boards (2xserial, 1xparallel, 2xIDE on
single board).

Maybe the buffered UARTs were common later? I moved away from modems to
ISDN (ISA card) around 1996 or 1997, then ADSL (PCI card) and then
mostly ethernet for communication.

On Thu, 23 Jun 2022 17:28:49 +0300
Anssi Saari <as@sci.fi> wrote:

> Maybe the buffered UARTs were common later? I moved away from modems to
> ISDN (ISA card) around 1996 or 1997, then ADSL (PCI card) and then
> mostly ethernet for communication.

They started to get common around 1994/5 - along with the rise in
fast modem use.

--
Steve O'Hara-Smith
Odds and Ends at http://www.sohara.org/

Anssi Saari <as@sci.fi> writes:
> Did they? I remember buying an ISA card by SIIG with the 16550 UART that
> had the huge 16 byte buffer back in 1993 when I got a 14400 bps
> modem. With the V.42bis compression I think you could have at least 2x
> compression so double data rate over the serial link. Might've had a 486
> by then though.
>
> In my computers the built in ports had just the no buffer
> 16450s. Same with the multi-I/O boards (2xserial, 1xparallel, 2xIDE on
> single board).
>
> Maybe the buffered UARTs were common later? I moved away from modems to
> ISDN (ISA card) around 1996 or 1997, then ADSL (PCI card) and then
> mostly ethernet for communication.

1993, had left IBM ... and was doing work from home. I got offer to do
modem drivers (unix, windows, dos) for PAGESAT in return for downlink
with full netnews feed. I had a RS6000/320 and SGI Indy on my desk and
couple 486 machines (one running waffle, bulletin board software that I
made the netnews feed available on) ... had 16550 UART boards. Also
wrote article on modem drivers & pagesat for boardwatch magazine (had a
picture of me in backyard with PAGESAT dish). Started out 9600, but they
had to double it to 19.2 with increase in images (and there were
periodic further increases).

16550
https://en.wikipedia.org/wiki/16550_UART
Pagesat
http://www.art.net/lile/pagesat/netnews.html
Pagesat at 115.2kbps
http://www.art.net/lile/ncit/service.html
Boardwatch
https://en.wikipedia.org/wiki/Boardwatch
Waffle
https://en.wikipedia.org/wiki/Waffle_(BBS_software)

--
virtualization experience starting Jan1968, online at home since Mar1970

To: Ahem A Rivet's Shot
-=> Ahem A Rivet's Shot wrote to alt.folklore.computers <=-

AAS> Then the 16550A came out with the bugs fixed. With them on two
AAS> 16 port cards I managed to run 32 modems at 19200 on a 50MHz 486
AAS> without dropping a byte.

Digiboards! Those things were awesome.

I ran a store POS system back in the day running Excellenet WAN software.
the hub ran on a 486 running OS/2, the store cash registers either ran OS/2
or DOS. Every night as the stores closed they'd shut down the registers,
close out the tills, and the register would "phone home", sending store
sales, credit card sales info, employee payroll and inventory figures, and
the hub would send back data updates and new price lookup files for the
registers.

I was running a dial-up BBS back then, the store-and-forward topology was
reminiscent of Fidonet. I'd often thought you could do the same thing with a
Fido mailer and save a ton of money.

The batch files running the stores were written in DOS Batch or REXX,
meaning if you fat-fingered an update, you were walking a non-technical
store manager through using a line editor over the phone to make the fix.

At certain times of the day, I'd have 32 14.4 multitech modems lit and
transferring files, and it worked like a charm.

kurt weiske | kweiske at realitycheckbbs dot org
| http://realitycheckbbs.org
| 1:218/700@fidonet

--- MultiMail/DOS v0.52
--- Synchronet 3.19c-Win32 NewsLink 1.113
* realitycheckBBS - Aptos, CA - telnet://realitycheckbbs.org

On Sat, 25 Jun 2022 09:50:00 -0700
"Kurt Weiske" <kurt.weiske@realitycheckbbs.org.remove-tq0-this> wrote:

> To: Ahem A Rivet's Shot
> -=> Ahem A Rivet's Shot wrote to alt.folklore.computers <=-
>
> AAS> Then the 16550A came out with the bugs fixed. With them on
> AAS> two 16 port cards I managed to run 32 modems at 19200 on a 50MHz 486
> AAS> without dropping a byte.
>
> Digiboards! Those things were awesome.

Correct, they were indeed Digiboards.

> At certain times of the day, I'd have 32 14.4 multitech modems lit and
> transferring files, and it worked like a charm.

It happened at 6pm on the dot every day for me, when the evening
cheap phone rate cut in - I was running a dial-up ISP.

I was once in Demon's Finchley office at 6pm and got to see their
bank of modems (90 or so at the time IIRC all running on a Sun box with
SCSI attached serial[1] ports) suddenly light up in a ripple across the
rack right on the dot of 6pm.

[1] Of course these days we have serial attached SCSI.

--
Steve O'Hara-Smith
Odds and Ends at http://www.sohara.org/

Ahem A Rivet's Shot <steveo@eircom.net> writes:
> [1] Of course these days we have serial attached SCSI.

trivia: circa 1990 , IBM Hursley did 9333, SCSI disk drives ... with
serial copper running packetized SCSI protocol, originally was
80mbits/sec full duplex (160mbits/sec aggregate). 1988, I was asked to
work with LLNL on standardizing some of serial stuff they were playing
with, quickly becomes fibre channel standard (initially 1gbit, full
duplex, 2gbit, 200GBYTE) I had hoped that 9333 could morpth into
interoperable 1/8 & 1/4 speed FCS, but instead it morphs into SSA
https://en.wikipedia.org/wiki/Serial_Storage_Architecture

--
virtualization experience starting Jan1968, online at home since Mar1970