9 Jun
2025
9 Jun
'25
5:57 p.m.
Hi,
I got interested in UI design and often study some historical aspects of it as I work on
software. It’s hard not to notice how fast/usable Text User Interfaces are—ncurses and its
siblings are still alive and well. From the ergonomy point of view, not needing a mouse in
those interfaces if perfect.
Question: where did TUIs come from originally, and what were their earliest instances?
Many pages state that Vi was the first, but I’ve been looking through some old hardware
photos, and things capable of more sophisticated interactions existed before Vi:
https://en.wikipedia.org/wiki/Light_pen
Some terminals with block display:
https://en.wikipedia.org/wiki/IBM_3270
^ ’71. Wiki says Vi showed up in ’76, but I suspect IBM mainframes may have had TUIs
before.
Question 2: were there any manuals talking about TUIs? I’m thinking some of those spiffy
IBM things mandating certain design.
Thanks,
Adam
9 Jun
9 Jun
6:21 p.m.
On Mon, 9 Jun 2025 at 14:14, Adam Koszek <adam(a)koszek.com> wrote:
Hi,
I got interested in UI design and often study some historical aspects of
it as I work on software. It’s hard not to notice how fast/usable Text User
Interfaces are—ncurses and its siblings are still alive and well. From the
ergonomy point of view, not needing a mouse in those interfaces if perfect.
Question: where did TUIs come from originally, and what were their
earliest instances?
Many pages state that Vi was the first, but I’ve been looking through some
old hardware photos, and things capable of more sophisticated interactions
existed before Vi:
https://en.wikipedia.org/wiki/Light_pen
Some terminals with block display:
https://en.wikipedia.org/wiki/IBM_3270
^ ’71. Wiki says Vi showed up in ’76, but I suspect IBM mainframes may
have had TUIs before.
Question 2: were there any manuals talking about TUIs? I’m thinking some
of those spiffy IBM things mandating certain design.
Does this count? I was just looking at it the other day.
https://en.wikipedia.org/wiki/Hypertext_Editing_System
I have a feeling we're going to get away from UNIX pretty quickly here.
-Henry
11 Jun
11 Jun
6:38 p.m.
I think it counts! I was suspecting TUIs were either an IBM thing or UNIX thing—not sure
if it’s < 1970 direction or > 1970 direction. In UNIX, someone must have added code
for the cursor addressing for CRT screens b/c on printer terminals moving back a page …
wasn’t possible?
Adam
On Jun 9, 2025, at 11:21 AM, Henry Bent
<henry.r.bent(a)gmail.com> wrote:
On Mon, 9 Jun 2025 at 14:14, Adam Koszek <adam(a)koszek.com
<mailto:adam@koszek.com>> wrote:
Hi,
I got interested in UI design and often study some historical aspects of it as I work on
software. It’s hard not to notice how fast/usable Text User Interfaces are—ncurses and its
siblings are still alive and well. From the ergonomy point of view, not needing a mouse in
those interfaces if perfect.
Question: where did TUIs come from originally, and what were their earliest instances?
Many pages state that Vi was the first, but I’ve been looking through some old hardware
photos, and things capable of more sophisticated interactions existed before Vi:
https://en.wikipedia.org/wiki/Light_pen
Some terminals with block display:
https://en.wikipedia.org/wiki/IBM_3270
^ ’71. Wiki says Vi showed up in ’76, but I suspect IBM mainframes may have had TUIs
before.
Question 2: were there any manuals talking about TUIs? I’m thinking some of those spiffy
IBM things mandating certain design.
Does this count? I was just looking at it the other day.
https://en.wikipedia.org/wiki/Hypertext_Editing_System
I have a feeling we're going to get away from UNIX pretty quickly here.
-Henry 
6:46 p.m.
ECMA 48 was first published in 1976 as a standard for terminal escape
sequences. This could be in support of any multiuser system, not just IBM
or UNIX.
Rik
On Wed, Jun 11, 2025 at 11:39 AM Adam Koszek <adam(a)koszek.com> wrote:
I think it counts! I was suspecting TUIs were either
an IBM thing or UNIX
thing—not sure if it’s < 1970 direction or > 1970 direction. In UNIX,
someone must have added code for the cursor addressing for CRT screens b/c
on printer terminals moving back a page … wasn’t possible?
Adam
On Jun 9, 2025, at 11:21 AM, Henry Bent <henry.r.bent(a)gmail.com> wrote:
On Mon, 9 Jun 2025 at 14:14, Adam Koszek <adam(a)koszek.com> wrote:
Hi,
I got interested in UI design and often study some historical aspects of
it as I work on software. It’s hard not to notice how fast/usable Text User
Interfaces are—ncurses and its siblings are still alive and well. From the
ergonomy point of view, not needing a mouse in those interfaces if perfect.
Question: where did TUIs come from originally, and what were their
earliest instances?
Many pages state that Vi was the first, but I’ve been looking through
some old hardware photos, and things capable of more sophisticated
interactions existed before Vi:
https://en.wikipedia.org/wiki/Light_pen
Some terminals with block display:
https://en.wikipedia.org/wiki/IBM_3270
^ ’71. Wiki says Vi showed up in ’76, but I suspect IBM mainframes may
have had TUIs before.
Question 2: were there any manuals talking about TUIs? I’m thinking some
of those spiffy IBM things mandating certain design.
Does this count? I was just looking at it the other day.
https://en.wikipedia.org/wiki/Hypertext_Editing_System
I have a feeling we're going to get away from UNIX pretty quickly here.
-Henry
10:53 a.m.
Besides all the curses and termcap/terminfo efforts, There was also POP
which stood for Pieces of Perspect. It was a terminal screen update library.
I think it used curses as the underlying CRT management package.
Sorry, i dont recall the authors. It was on one of the distribution tapes
around the late 1970s, early 80s. I remember it being painfully slow on my
PDP 11/70. It would redraw the entire screen with every change. I sped it
up using differential compute buffers so only actual changes were refreshed.
I used it for my two player Tank War game and a new email program I called
Mr. ZIP. Mr. Zip had three display zones or display panels. One for email
actions. One for a list of emails and a email contents zone for displaying
the text of the selected email. Part of the work I did hacking the kernel,
utilities and applications at NJSMU so I could support 250 users connected
via serial to the PDP 11/70 used to run that insurance company.
Massive 4 MB of National Semiconductor Ram, The maximum memory for that
machine.
In those days of Tape based distribution, My boss was not keen to engage
in "software distribution". More out of fear of running afoul of AT&T
lawyers, I think than anything else.
So releasing source code was forbidden. Otherwise, POP might have taken
off with my performance improvements. You would have all been enthralled by
Tank War and got no real work done. You had to enter several of your moves
ahead of time. (Command buffering)
So you really had to think strategically to out maneuver your opponent.
On Mon, Jun 9, 2025, 1:57 PM Adam Koszek <adam(a)koszek.com> wrote:
Hi,
I got interested in UI design and often study some historical aspects of
it as I work on software. It’s hard not to notice how fast/usable Text User
Interfaces are—ncurses and its siblings are still alive and well. From the
ergonomy point of view, not needing a mouse in those interfaces if perfect.
Question: where did TUIs come from originally, and what were their
earliest instances?
Many pages state that Vi was the first, but I’ve been looking through some
old hardware photos, and things capable of more sophisticated interactions
existed before Vi:
https://en.wikipedia.org/wiki/Light_pen
Some terminals with block display:
https://en.wikipedia.org/wiki/IBM_3270
^ ’71. Wiki says Vi showed up in ’76, but I suspect IBM mainframes may
have had TUIs before.
Question 2: were there any manuals talking about TUIs? I’m thinking some
of those spiffy IBM things mandating certain design.
Thanks,
Adam
9:46 p.m.
At Mon, 9 Jun 2025 10:57:05 -0700, Adam Koszek <adam(a)koszek.com> wrote:
Subject: [TUHS] Where/when did TUIs come from
Question: where did TUIs come from originally, and what were their
earliest instances?
By "TUI", I assume you mean that in the anachronistic sense going
backwards from "GUI". (Technically command-line interfaces are "text
UIs" too, but just using one line at a time!)
So "TUI" is really "full-screen text-based user interface"?
I would think any terminal "block oriented" terminal, including all the
way back to the IBM 2260 (1964), were typically used to create
"full-screen textual user interfaces", while character-oriented
terminals, including printing terminals, were typically used to create
command-line interfaces (at least until screen-oriented terminals
supported faster baud rates).
I would also think the first full-screen editors (on any system,
including with any kind of character-oriented terminals) should arguably
be described as "TUIs" (as opposed to line-oriented editors and other
command-line interfaces).
In the Unix world I think there were versions of full-screen editors
well before "vi" was fully fledged. I used one called "fred -- the
friendly editor", initially IIRC on a PDP 11/60 running 7th Edition Unix
at the University of Calgary, though at the time it was contemporary
with "vi" (1980). It was, I believe, basically also an adaptation of
Unix "ed". It had a modeless "open mode" controlled by the cursor
keys
allow one to move about and change text at arbitrary locations on the
screen, though it could also be used in a more line-oriented way like
you would expect a full-screen "ed" to work -- in this mode it still
maintained a full screen's worth of context and one could "page" up and
down in the document being edited but typically one inserted/appended
lines in a modal manner, or moved or deleted or ran other commands on a
line or block of lines at a time.
Unfortunately I've never seen or heard any more of "fred" since leaving
UofC in the early 1980s. It may even have been a local name for some
adaptation of another editor, perhaps even "em" itself.
In second year I quickly quit using "fred" and switched to (Gosling)
emacs (true full-time modeless operation, finally!) as soon as it was
available on the new VAX 11/780 (then running 4BSD -- though in first
year it initially ran 32V then 3BSD).
--
Greg A. Woods <gwoods(a)acm.org>
Kelowna, BC +1 250 762-7675 RoboHack <woods(a)robohack.ca>
Planix, Inc. <woods(a)planix.com> Avoncote Farms <woods(a)avoncote.ca>
12 Jun
12 Jun
3:06 p.m.
On Wed, Jun 11, 2025 at 5:46 PM Greg A. Woods <woods(a)robohack.ca> wrote:
In the Unix world I think there were versions of full-screen editors
well before "vi" was fully fledged. I used one called "fred -- the
friendly editor", initially IIRC on a PDP 11/60 running 7th Edition Unix
at the University of Calgary, though at the time it was contemporary
with "vi" (1980). It was, I believe, basically also an adaptation of
Unix "ed". It had a modeless "open mode" controlled by the cursor
keys
allow one to move about and change text at arbitrary locations on the
screen, though it could also be used in a more line-oriented way like
you would expect a full-screen "ed" to work -- in this mode it still
maintained a full screen's worth of context and one could "page" up and
down in the document being edited but typically one inserted/appended
lines in a modal manner, or moved or deleted or ran other commands on a
line or block of lines at a time.
fred was original released for v6. FWIW: I think it came out of Cornell. I
know we had it at CMU in 1977/78 and ran it on Perkin Elmer Fox terminals
mostly, since that is what most of the UNIX boxes we had. I'm pretty sure
that I have the sources somewhere. There was nothing like termcap; it was
hard-coded to the terminal being used in some manner, probably reading
something static like /etc/ttys and knowing the site's set up. But as you
said, it was small enough to run on 11/40 class UNIX systems which helpful
since 40s and 34s were what we had the most. We did not have vi nor emacs
in UNIX-land which was still PDP-11's (EMACS was on the PDP-10s).
13 Jun
13 Jun
1:42 p.m.
Clem Cole <clemc(a)ccc.com> writes:
fred was original released for v6. FWIW: I think it
came out of Cornell. I
know we had it at CMU in 1977/78 and ran it on Perkin Elmer Fox terminals
mostly, since that is what most of the UNIX boxes we had.
Another one from around the same time is 'ded'. I use it under PWB 1.0
on two of my 11/23 systems, and quite like it. There's a link to an
article about it, and where to find it in the TUHS archives, on my web
page at https://hamartun.priv.no/ded.html
-tih
--
The creation of the state of Israel was a regrettable mistake. It is
time to undo this mistake, and finally re-establish a free Palestine.
16 Jun
16 Jun
10:51 p.m.
New subject: screen editors, Where/when did TUIs come from
It appears that Greg A. Woods <tuhs(a)tuhs.org> said:
In the Unix world I think there were versions of
full-screen editors
well before "vi" was fully fledged. ...
At Yale we had one called "e" that ran on our early bitmap terminals. Although
programs could directly write arbitrary graphics to them, the editor used them
as text terminals via the system terminal emulator I wrote. I made a few tweaks
to help the editor like insert mode to push characters out of the way rather than
having to redraw the line each time, and an input mode that buffered up characters
until you typed something that needed more than an echo to the screen.
It was based on Ned Irons' PDP-10 screen editor that ran on Omron glass ttys,
and was a cousin of INed which Walt Bilofsky wrote for Interactive Systems.
They used the quarter-plane model, you could move the cursor around the screen
and overtype stuff, with no automatic flow from one line to another. There were
control characters to do the other editing stuff like search and save/exit, but
nowhere near as complicated as emacs.
R's,
John
17 Jun
17 Jun
12:48 p.m.
New subject: screen editors, Where/when did TUIs come from
John Levine wrote:
At Yale we had one called "e" that ran on
our early bitmap terminals.
The GEM display system, in case anyone is curious. Here's a description:
https://engineering.yale.edu/application/files/3917/3714/8796/tr163.pdf
Unrelated to DR's GEM desktop.
It was based on Ned Irons' PDP-10 screen editor
that ran on Omron glass ttys,
and was a cousin of INed which Walt Bilofsky wrote for Interactive Systems.
I'm curious if any listing or simliar may have survived for these?
12:52 p.m.
New subject: screen editors, Where/when did TUIs come from
John Levine wrote:
> At Yale we had one called "e" that ran on our early bitmap terminals.
Wasn't this related (somehow) to what was called the RAND editor?
I remember using "e" on IS/1, a V6 derivative on the PDP-11, in
the 1980-81 time frame. I don't know if it was hardwired for a
particular terminal or not. The ones we had at my school had some
smarts in them, you could do simple bitmap graphics. Someone wrote
a "Space Invaders" game for them. I don't remember what brand they
were.
Arnold
3:06 p.m.
New subject: screen editors, Where/when did TUIs come from
On Tue, 17 Jun 2025, arnold(a)skeeve.com wrote:
Yes. The same people went from Yale to RAND and then started Interactive.
INed may have used code from the RAND editor, even if not, it was written
by the same guy and worked much the same.
> At Yale we
had one called "e" that ran on our early bitmap terminals.
Wasn't this related (somehow) to what was called the RAND editor? I remember using "e" on IS/1, a V6
derivative on the PDP-11, in
the 1980-81 time frame. I don't know if it was hardwired for a
particular terminal or not.
Interactive sold OEM terminals but I don't recall whether the editor also
worked on other devices.
Regards,
John Levine, johnl(a)taugh.com, Taughannock Networks, Trumansburg NY
Please consider the environment before reading this e-mail. https://jl.ly
3:40 p.m.
New subject: screen editors, Where/when did TUIs come from
Somebody (ISC I >>think<<) inserted some of the code for INed into the ROMs
of the Perkin-Elmer Fox terminals to provide some assistance (which were
6800-based). I never knew much about it, but it was the system the 68000
folks (Les Crudele, Nich Trudenick et al) used during core development. It
all ran on a PDP-11/70, which was running Interactive's V6 port that their
group controlled in the back, basically unnoticed. Les told me that the
system was pretty responsive, even with the load they put on it of
editing and text processing, much more responsive than the resources
Motorola corporate offered them. This was all important because the 6809
was the "corporate" project and what would become the 68000 was purely an
unamed skunkworks project, basically hidden away from the core management
structure as an experiment/bet that they could do it (a story told
elsewhere and while UNIX rely on the results, was not part of UNIX's story
directly other than a PDP-11/70 running UNIX was the system was used to
create the processor).
On Tue, Jun 17, 2025 at 11:06 AM John R Levine <johnl(a)taugh.com> wrote:
On Tue, 17 Jun 2025, arnold(a)skeeve.com wrote:
Yes. The same people went from Yale to RAND and then started Interactive.
INed may have used code from the RAND editor, even if not, it was written
by the same guy and worked much the same.
> At
Yale we had one called "e" that ran on our early bitmap terminals.
Wasn't this related (somehow) to what was called the RAND editor? I remember using "e" on IS/1, a V6
derivative on the PDP-11, in
the 1980-81 time frame. I don't know if it was hardwired for a
particular terminal or not.
Interactive sold OEM terminals but I don't recall whether the editor also
worked on other devices.
Regards,
John Levine, johnl(a)taugh.com, Taughannock Networks, Trumansburg NY
Please consider the environment before reading this e-mail. https://jl.ly
4:27 p.m.
New subject: screen editors, Where/when did TUIs come from
INtext Terminal, INed and other ISC products ~ 1979 mentioned in
https://archive.org/details/TNM_Automated_office_communications_-_Interacti…
On Tue, Jun 17, 2025 at 11:40:14AM -0400, Clem Cole wrote:
Somebody (ISC I >>think<<) inserted some
of the code for INed into the ROMs
of the Perkin-Elmer Fox terminals to provide some assistance (which were
6800-based). I never knew much about it, but it was the system the 68000
folks (Les Crudele, Nich Trudenick et al) used during core development. It
all ran on a PDP-11/70, which was running Interactive's V6 port that their
group controlled in the back, basically unnoticed. Les told me that the
system was pretty responsive, even with the load they put on it of
editing and text processing, much more responsive than the resources
Motorola corporate offered them. This was all important because the 6809
was the "corporate" project and what would become the 68000 was purely an
unamed skunkworks project, basically hidden away from the core management
structure as an experiment/bet that they could do it (a story told
elsewhere and while UNIX rely on the results, was not part of UNIX's story
directly other than a PDP-11/70 running UNIX was the system was used to
create the processor).
On Tue, Jun 17, 2025 at 11:06 AM John R Levine <johnl(a)taugh.com> wrote:
> On Tue, 17 Jun 2025, arnold(a)skeeve.com wrote:
> >>> At Yale we had one called "e" that ran on our early bitmap
terminals.
> >
> > Wasn't this related (somehow) to what was called the RAND editor?
>
> Yes. The same people went from Yale to RAND and then started Interactive.
> INed may have used code from the RAND editor, even if not, it was written
> by the same guy and worked much the same.
>
> > I remember using "e" on IS/1, a V6 derivative on the PDP-11, in
> > the 1980-81 time frame. I don't know if it was hardwired for a
> > particular terminal or not.
>
> Interactive sold OEM terminals but I don't recall whether the editor also
> worked on other devices.
>
> Regards,
> John Levine, johnl(a)taugh.com, Taughannock Networks, Trumansburg NY
> Please consider the environment before reading this e-mail. https://jl.ly
>
4:29 p.m.
New subject: screen editors, Where/when did TUIs come from
On Tue, 17 Jun 2025, Clem Cole wrote:
Somebody (ISC I >>think<<) inserted some
of the code for INed into the ROMs
of the Perkin-Elmer Fox terminals to provide some assistance (which were
6800-based).
Yup, those were called INterms or something like that.
folks (Les Crudele, Nich Trudenick et al) used during
core development. It
all ran on a PDP-11/70, which was running Interactive's V6 port that their
We used whatever was available from AT&T so it'd have been v7 probably in
1979. Later it was Sys III.
Regards,
John Levine, johnl(a)taugh.com, Taughannock Networks, Trumansburg NY
Please consider the environment before reading this e-mail. https://jl.ly
2:21 p.m.
New subject: Yale and RAND editors
On Tue, Jun 17, 2025 at 12:48:19PM +0000, Lars Brinkhoff wrote:
Walter Bilofsky
The CRT Text Editor NED - Introduction and Reference Manual
https://apps.dtic.mil/sti/tr/pdf/ADA049314.pdf
has a page on the history of associated editors.
tuhs Applications/Usenix_77/ug091377.tar.gz
2/rand/re/
/* file re.c - main program for RAND editor */
tuhs Applications/Shoppa_Tapes/usenix_80_delaware.tar.gz
delaware/maryland/rand/re/
delaware/maryland/rand/redoc/
/* file ned.c - main program for new RAND editor */
/* Walt Bilofsky - 14 January 1977 */
John Levine wrote:
https://engineering.yale.edu/application/files/2117/3714/8646/TR19_The_Yale…
references an IDA CDC6600 editor
Edgar T. Irons, Frans M. Djorup
A CRT Editing System
https://dl.acm.org/doi/pdf/10.1145/361237.361244
At Yale we had one called "e" that ran
on our early bitmap terminals.
The GEM display system, in case anyone is curious. Here's a description:
https://engineering.yale.edu/application/files/3917/3714/8796/tr163.pdf
Unrelated to DR's GEM desktop. It was based on Ned Irons' PDP-10 screen
editor that ran on Omron glass ttys,
and was a cousin of INed which Walt Bilofsky wrote for Interactive Systems.
I'm curious if any listing or simliar may have survived for these?
3:11 p.m.
New subject: Yale and RAND editors
On Wed, 18 Jun 2025, Jonathan Gray wrote:
Huh, didn't realize that report was still available. It became an article
in Software Practice and Experience.
Regards,
John Levine, johnl(a)taugh.com, Taughannock Networks, Trumansburg NY
Please consider the environment before reading this e-mail. https://jl.ly
The GEM
display system, in case anyone is curious. Here's a description:
https://engineering.yale.edu/application/files/3917/3714/8796/tr163.pdf
Unrelated to DR's GEM desktop.
https://engineering.yale.edu/application/files/2117/3714/8646/TR19_The_Yale…
1
days inactive
9
days old
16 comments
12 participants
participants (12)
-
Adam Koszek -
arnold＠skeeve.com -
Clem Cole -
Greg A. Woods -
Henry Bent -
John Levine -
John R Levine
-
Jonathan Gray -
Kenneth Goodwin -
Lars Brinkhoff -
Rik Farrow -
Tom Ivar Helbekkmo