The Origin of Emacs in 1976
onlisp.co.uk
onlisp.co.uk
http://pdp-10.trailing-edge.com/mit_emacs_170_teco_1220/01/e...
From this we can conclude that humans do not need any nice things like programming languages, assemblers or compilers to program effectively.
It's a very readable language that anybody who knows Python etc could understand and has nothing to do with any of this.
EDIT: LOL, downvotes why?
http://probablyprogramming.com/2009/07/11/the-aplus-programm...
But maybe not efficiently. I realize that was tongue and cheek, but since this sometimes comes up in earnest (especially when it comes to dynamic vs. static type systems), it's worth noting that we knew from the very beginning that we can effectively program in what we'd now call adverse conditions: Not only were some earlier computers programmed directly in binary code (and Woz famously still programmed a 6502 by entering hex bytes into a monitor), before that we programmed by plugging in cords into panels, and in some sense still do in logic circuit design.
However, we sure became much more efficient in programming not only because higher level abstractions allow us to do things quicker, but because in addition of static type systems, we either catch really hard to debug issues much earlier, or won't make a lot of mistakes in the first place (you won't accidentally save a temporary into some register that you already used for something else even if it's "just C").
One nice thing about Typescript is that it gives us a mostly statically typed language and it has a control to compare against (JavaScript). That allows programmers to learn the strengths and weaknesses of both approaches. Previously most comparisons were apples versus oranges.
I'm guessing Python's Type hints (or Cython) also give a taste of the positives and negatives of static typing.
I would also presume both Typescript and Cython miss out on some of the more extreme optimisation benefits of static typing?
I'm not sure if I've got things quite compatible enough to run all of those macros. I could run them in my Virtual VAX/VMS system, and see how they do on TECO32 there.
To: CBF at MIT-AI, EAK at MIT-AI, ED at MIT-AI, MOON ay MIT-AI
I realize this was before DNS was invented, so I am not surprised by the lack of TLD.
Yes, “ at “ was still in use. ISTR typing ‘@‘ but this was a long time ago (I didn’t get to the AI lab until ‘78 by when EMACS was well entrenched).
This was pre TCP, pre RFC 733, pre DNS etc. I don’t remember if mail already had its own transport or if mail was still a special mode of the FTP protocol. I think that was still being argued over.
Here is a very early map of the routers.
https://groups.google.com/g/net.general/c/u1dxxNcTIX0/m/mHMW...
unfortunately 'show original message' is grayed out, so i can't actually see the map
<https://en.wikipedia.org/wiki/Incompatible_Timesharing_Syste...>
Of course, especially given its relative simplicity, at some point non-UNIX systems started to implement it (I think I maybe used some DOS package at some point).
_some poking around in Wikipedia_
Indeed, FTP is from 1971 and UUCP is from 1979. And the SMTP page describes that it all started as a mode in FTP.
https://github.com/larsbrinkhoff/mldev
>MLDEV
>The "ML Device" protocol is a remote file system invented for the PDP-10 MIT Incompatible Timesharing System. It may have been the very first transparent networking filesystem.
>Build: Install libfuse-dev, then type make.
>Usage: Create a mount point directory, and then mount an ITS file system like this:
>sudo ./mount.mldev <hostname> <mountpoint>
What are some early network file systems?
https://retrocomputing.stackexchange.com/questions/13709/wha...
>I know it's always hard to define an exact first, so I'll just ask about early network file systems. To be more specific, I'm wondering about file systems that transparently present directories and files as if they were local. I don't consider FTP or similar mechanisms to be candidates.
>For some time it has been my suspicion that the ITS "MLDEV" facility was very early ("the first"), but I never had any facts to back it up. I do have a file timestamped November 1972 that gives us an upper bound for the introduction of the facility. On the MIT-AI PDP-10, it provided an ML device (hence the name) which worked much like the local DSK device except the data was accessed through MIT's ARPANET IMP. MIT-ML of course had an AI device.
>Present evidence is that ITS had a very early networked file system no later than 1972.
>ITS has had a facility called the "ML Device", or MLDEV, for a long time. The name probably comes from a time when the users of the older more established computer called AI wanted access to files on the computer called ML. The "device" part is because ITS files are named by four parts: device, directory, and two file names. To access ML files from AI, the device name is ML. Conversely, from ML the AI device is used.
>To date this facility, we can check timestamps from backup tapes. The earliest file called MLDEV is from July 1975. However, if we examine the source code for an earlier ITS version we see that it has built in ML and AI devices that when accessed starts a program called SYS; ATSIGN OTHER. Disassembling this file we can see that it indeed is accepting file operations and sends information across ARPANET. The earliest file with that name has a timestamp from November 1972.
>We know from RFC documents 342, 344, and 366 that AI and ML were first put on the ARPANET in May-June 1972. So MLDEV could not have been in use before 1972.
List of Chaos application protocols:
https://gunkies.org/wiki/List_of_Chaos_application_protocols
Among the many hackerish features of ITS was allowing a user space process to act like a device, not just for a filesystem but as a TTY, which allowed a lot more kinds of terminals to be supported than would fit in the monitor (kernel in today’s parlance).
All in the early 70s.
But even without CRTSTY, ITS gave everyone the ability to reattach to their processes after they got disconnected, and even pass processes like Lisp interpreters and Zork games around from user to user to share like a joint, by detatching and reattaching them to different TTYs!
https://github.com/PDP-10/its/blob/master/doc/_info_/its.pri...
>VIII.B. CRTSTY
>For display terminals which are not handled directly by the system there exists a program called CRTSTY. This provides the software support to permit these terminals to take advantage of their display characteristics. Among the terminals supported by CRTSTY are: ADM3's, Concept 100's, Hazeltine 1500's, Perkins Elmer FOX's, Datamedias, Control Data Corporation 713's, and a host of others. You should try :INFO CRTSTY and :INFO TERMS to see if your terminal is included. If it isn't, the maintainers of CRTSTY are usually willing to help you add it. In the CRTSTY section of INFO the information you must provide to have a new terminal supported is listed. You can communicate with the maintainers of CRTSTY by sending mail to BUG-CRTSTY.
https://github.com/PDP-10/its/blob/master/doc/humor/jargon.6...
>SUPDUP v. To communicate with another ARPAnet host using the SUPDUP program, which is a SUPer-DUPer TELNET talking a special display protocol used mostly in talking to ITS sites. Sometimes abbreviated to SD.
https://datatracker.ietf.org/doc/rfc734/
>NWG/RFC# 734 MRC 07-OCT-77 08:46 41953 SUPDUP Display Protocol Page 1 Network Working Group Mark Crispin Request for Comments 734 SU-AI NIC 41953 7 October 1977
>SUPDUP Protocol
>This document describes the SUPDUP protocol, a highly efficient display telnet protocol. It originally started as a private protocol between the ITS systems at MIT to allow a user at any one of these systems to use one of the others as a display. At the current writing, SUPDUP user programs also exist for Data Disc and Datamedia displays at SU-AI and for Datamedias at SRI-KL. The author is not aware of any SUPDUP servers other than at the four MIT ITS sites.
>The advantage of the SUPDUP protocol over an individual terminal's protocol is that SUPDUP defines a "virtual" or "software" display terminal that implements relevant cursor motion operations. The protocol is not built on any particular display terminal but rather on the set of functions common to all display terminals; hence it is completely device- independent. In addition, the protocol also provides for terminals which cannot handle certain operations, such as line or character insert/delete. In fact, it is more than this. It provides for terminals which are missing any set of features, all the way down to model 33 Teletypes.
>The advantage over the TELNET protocol is that SUPDUP takes advantage of the full capabilities of display terminals, although it also has the ability to run printing terminals.
https://its.victor.se/wiki/_media/ai_wp_235.pdf
>The SUPDUP protocol [Crispin 77] is a highly efficient display telnet protocol. The advantage over the TELNET protocol is that SUPDUP takes advantage of the full capabilities of display terminals, although it also has the ability to run printing terminals. When you use the SUPDUP protocol, you do not need to tell the remote host which you are connecting to what type of terminal you have or what the terminal's capabilities are. The host you are SUPDUPing from handles the actual display support for your terminal.
>Additionally, SUPDUP defines a network graphics protpcol [Stallman 78] which makes it easy for network hosts to draw pictures along in addition to text.
SUPDUP Graphics Extension:
https://github.com/PDP-10/its/blob/master/doc/sysdoc/grphcs....
You didn't need CRTSTY to implement the SUPDUP protocol, but CRTSTY translated SUPDUP to standard terminal escape sequences, and provided buffering and optimization. Lisp machines and Knight TVs had their own SUPDUP terminal programs, and I made one for my Apple ][ in Forth.
SUPDUP had a line saving extension that RMS hacked into ITS Emacs, so it can stash lines on the screen in local memory before overpainting them, then almost instantly pop them back on the screen later, so you can scroll back and forth through text really fast at 300 baud (plus Devon McCullough made ZipMod with Huffman encoding to make it even faster on top of that):
https://news.ycombinator.com/item?id=15539983
I wrote a terminal emulator for the Apple ][ + Videx 80 column card in FORTH and reverse polish notation 6502 code, that supported SUPDUP including the line saving protocol that RMS hacked into ITS and EMACS (%TDSAV %TDRES), so Emacs could stash lines on the screen in a memory buffer (on the Apple's bank switched RAM card) and restore them later to paint the screen quickly when scrolling back and forth. (That was great at 300 baud!)
This is just the part of it that interprets SUPDUP codes -- layered on top of the rest of the terminal emulator that emulated H19 (VT52 + character/line insert/delete):
http://www.donhopkins.com/home/archive/forth/supdup.f
https://news.ycombinator.com/item?id=37973838
At 300 baud over a congested ARPA connection, we had a lot of time to think about how to optimize terminal emulation while waiting for the screen to repaint...
When James Gosling later implemented his Unix version of Emacs, he created a super-hot screen update algorithm that automatically took into account the capabilities of terminals and the length of their escape codes and padding required (some slow terminals needed padding when performing line/character insert/delete operations, otherwise they would miss characters. It used dynamic programming to figure out the optimal way of updating the screen, which was really great at low baud rates like 300 or 1200 baud.
Read the paper and display.c code I linked to in the above linked comment, and see the wikipedia page on dynamic programming:
>James Gosling's Emacs screen redisplay algorithm also used similar "dynamic programming techniques" to compute the minimal cost path through a cost matrix of string edit operations (the costs depended i.e. on the number of characters to draw, length of the escape codes to insert/delete lines/characters, padding for slow terminals, etc).
[...]
https://news.ycombinator.com/item?id=37974319
The comments in this old PDP-10 SUPDUP code [5] are hilarious!
SUPDUP on ITS was so smart and opinionated, it had its own ideas about who was a loser and who was a winner, how to figure it out by checking INQUIR entries and relationships, and who it should never or always hassle about net-hopping (hopping through a PDP-10 over the net just to use it as a glorified terminal emulator).
It never hassled gumby or cstacy (because they were free to do whatever they wanted to, like flushing whiny annoying over-entitled alcoholic science fiction writers [1] [2] [3] [4]), but always hassled rms (because it was usually an imposter logging in as rms with password rms just to net hop through the system.)
[...]
;;; these people aren't hassled
wintab: irp x,,[gumby,alan,cstacy,cent,devon,dph,oaf,mly,jtw,jnc,klotz,cbf,sra,ed,swa,map,jbvb]
sixbit/x/
termin
winlen==.-wintab
;;; These people are always hassled
lostab: irp x,,[tk,hewitt,rms]
sixbit/x/
termin
loslen==.-lostab
https://news.ycombinator.com/item?id=37974319
[1] http://www.stormtiger.org/bob/humor/pournell/story.html
[2] https://news.ycombinator.com/item?id=15206151
[3] https://news.ycombinator.com/item?id=37723853
[4] https://news.ycombinator.com/item?id=37780515
[5] https://github.com/PDP-10/its/blob/master/src/sysnet/supdup....
I was able to unplug the USB-serial adapter from the embedded device's USB port. Then plug in a different USB-serial adapter (different chip requiring a different driver). It would obtain the same /dev/ttyUSB<n> device, and you would be back in your session intact. If you were in the middle of a vi edit, you could just Ctrl-L to refresh the screen and keep going.
I begin to understand where Plan 9 got that "other" kind of inspiration.
Speaking of intercept characters, did you know Will Crowther, one of the authors of Adventure, wrote the "Users Guide to the Terminal IMP"? You're in a maze of twisty little IMPs!
https://apps.dtic.mil/sti/tr/pdf/ADA014398.pdf
Don Woods, the other author of Adventure, is a great PostScript programmer! I worked with him at Sun on NeWS, but earlier he was DON@AI, so occasionally he'd get weird email intended for me.
Here's his Spider card game for NeWS, entirely written in PostScript:
https://www.donhopkins.com/home/archive/NeWS/news-tape/fun/s...
Maybe you could convert it to run in a laser printer, but it would sure use a lot of paper!
I don’t remember either. Could be an artifact of all the file translation over the years, or even early cruddy OCR.
None of the folks in that list would have been using TIPs — ^R mode was added after KTVs were deployed.
Two more curiosities:
- The sender’s address uses two at-signs, like NAME@@DOMAIN.
- There are typos where “at” is misspelled “ay”. Glad we don’t have to deal with that.
Ah yes, you'd see those coming out of the ARPANET site run by Arthur Fonzarelli. He also used to moderate rec.motorcycles.
Looking at https://datatracker.ietf.org/doc/html/rfc733 @ was already used behind the scenes in its current meaning, we're probably seeing a nicely human readable version of the emails sent.
Note I did not have the time to read the whole RFC and the above may be wrong.
In addition to what gumby wrote, CSAIL still maintains a `finger` server.
We suggest that the text of network mail, whether transmitted over the FTP telnet connection (via the MAIL command) or over a separate data connection (with the MLFL command), be governed by the syntax below:
Example:
From: White at SRI-ARC
Date: 24 JUL 1973 1527-PDT
Subject: Multi-Site Journal Meeting Announcement
NIC: 17996
At 10 AM Wednesday 25-JULY there will be a meeting
to discuss a Multi-Site Journal in the context of
the Utility. Y'all be here.
RFC 733 (Standard for the Format of ARPA Network Text Messages, 1977) has examples of addresses in both formats and this note: Whenever a message might be transmitted or migrate to a host on another network, full hierarchical addresses must be specified. These are indicated as a series of words, separated by at-sign or "at" indications.
Interesting. I thought Ameria preferred the "M-D-Y" order?
Americans colloquially prefer m/d/yy (slashes) rather than dashes, anyway.
: dev3; cat longnow.c
#include <stdio.h>
int
main()
{
long then = 0;
long long now = 0;
long long long time = 0;
printf("then was %ld, now is %lld, time is %llld\n", then, now, time);
return 0;
}
: dev3; gcc longnow.c
longnow.c: In function ‘main’:
longnow.c:8:13: error: ‘long long long’ is too long for GCC
8 | long long long time = 0;
| ^~~~I don't know how it's future proof, but that's what I've heard.
But hey, everybody has their small idiosyncratic pet peeves: I've ran twice into the same person here on HN who just can't stand term "x64" because it's "unofficial", but somehow refuses to use e.g. "EM64T" for it as well.
When it comes to this, the gods of Y100,000 will simply wish it with an act of will, and it will come about.
"Home was a couple of thousand light years away, which distance they traveled instantaneously, by an application of will. " -Joe Haldeman, "The Groombridge Effect"
"In 02024, Danny Hillis answered frequently asked questions about The Clock."
Interesting to see '76 email formats had To: and CC: but not Subject: and Date: info was rolled into what now we might call the Envelope, projected into the header.
I used EMAS and Vax/VMS based email in '79 which was sufficiently second-system to have a standard to implement against and Type:Value was in, with Date: and Subject:
And to be more extreme: you could actually write small assembly routines and run them in the registers (the first 20 addresses in the PDP-10 address space were the registers — super fast bc they were implemented with semiconductors rather than core!). That was hardcore and I only did it once or twice, just for fun.
Would you want to run the local assembly language (x86 or ARM) or are you thinking of also doing this with a PDP-10 emulator?
Either way, this would be rather cool.
Line editors are much less of a problem. In these I typically did not insert text in the middle of a line, I simply changed/substituted text in the middle of the line. This helps maintain context in contrast to moving a point around.
Fortunately, my TECO exposure was mostly tempered through VTedit, a screen oriented macro set for TECO on the PDP. Couldn’t escape it completely, but it helped reduce the load in my challenged, tiny brain.
By the time I moved to VMS, they had switched to EDT. Much more accessible.
SOS led naturally to ed in due course. I did learn emacs, But never really took to it. Now, I run my daily work log in org mode and do nothing else in it, but it is very good as a massively over-engineered multi-window table editor.
Charles Forsyth (York) wrote a macro for ed which took the point, redrew the 24 line vt00 display to show the lines around it (12 above 11 below) and put you back where you were. He said it made ed as good as vi, and on the baudrate of the day who am I to disagree?
The other nice feature was that you could do quick in place edits from the command line. Combination of ed and sed.
Another aspect that impacted TECO on the PDP was the concept of paging the file in. The machine was limited enough that you as a user were responsible for swapping parts of the buffer in and out, which, while usable, was not a good user experience.
When loading a large file TECO/vtedit would prompt “loading file xxx.yyy slowly”. And it was. In the labs, my friend and I would commandeer multiple terminals, with one of them being our file loaded in the editor. We’d constantly write the file out, but never exit.
And by “large file”, I’m talking a 35K byte source file.
It was a different time, to be sure.
Text would appear character by character, so it would take a good few seconds to show 24(!) lines of code.
At Waikato university a student, Ivan Piacun, wrote "vted", also a set of macros for TECO, on the PDP-11/70. I would say it was what most people used until the switch to VAX.
I tracked Ivan down on LinkedIn a few months ago. He says when he wrote "vted" he had absolutely no knowledge of other similar efforts, or of either emacs or vi, both of which existed (on the other side of the world) a couple of years earlier.
Everyone worked in a silo in those days.
You might eventually see something in BYTE (though it focussed on microcomputers) or Dr Dobbs (ditto) or Wireless World. Or something might make it ono a DECUS tape. But mostly you were just completely isolated if you lived on an island in the South Pacific Ocean, as we did.
I was happy with EDT, never got to like TPU's "Eve", though I didn't use it for long because I graduated ...
... graduated to a company that used DG MV10000 with AOS/VS, which had an even worse clone of TECO called SPEED.
There must have been a better editor, though I think still a line editor, but I have absolutely no memory of what it was called or how it worked!! I remember having to use SPEED for some things, but not most of the time. Maybe on very large files.
I was writing programs in PL/I and also Postscript on that machine. I hooked the brand new Apple Laserwriter to the MV10000, something both DG and the Apple dealer said was impossible, and used it, at first, mostly to replace an HP pen plotter to make diagrams and graphs to physically paste into gaps left in hard copy "camera ready" output from an IBM DisplayWriter word processing system.
The full-screen editor was SED, which I remember as being a bit less odd than vi, but maybe because I encountered it first.
http://www.simulogics.com/archive/public/docs/dg/sw/os/aos/0...
http://www.simulogics.com/archive/public/docs/dg/sw/os/aos/0...
And dozens and dozens of other old DG manuals.
And an online Nova simulator, and binaries for downloadable DG simulators apparently based on SimH (which I thought did PDP/ VAX).
Back in those days, csh was more popular as an interactive shell. sh was used more for scripting. At least that's what I recall. bash didn't exist yet.
From monochrome terminals to the latest (Android at least) phones you can still pull up a terminal and edit like it's 1984. Even on Windows with Cygwin.
If you had to bet on a technology still being there 40 years from now it would surely be that.
more generally, though, even if you aren't using a gui library with a built-in text editor widget, gigahertz and gigaflops make it not especially challenging to write a usable text editor in a high-level language. http://canonical.org/~kragen/sw/dev3/edumb.py is 183 lines of fairly straightforward python and includes full-text search and infinite undo; it's not quite usable but pretty close. its really stupid buffer data structure might start to get unresponsive once you're editing a file in the neighborhood of 100k
arguably gnu emacs is almost like multics emacs in this sense; the editing functionality of gnu emacs without any lisp code loaded is not actually zero but is pretty minimal. and i was surprised to find the other day that gnu emacs on my laptop is now compiling all my elisp to machine code, which is another similarity to maclisp, albeit a very recent one
I also had to do a little bit of magic to get hyper recognised as a modifier.
It would be awesome if someone could get Hyper added to the official list of USB keyboard keys.
at the other end of the spectrum, i just hacked together this stupid shell script to blindly reimpose my desired xkb configuration with lower-case parens yesterday: http://canonical.org/~kragen/sw/dev3/kbdd maybe you know the actually correct way to do this?
Sure could help out .. pipe me an email. It isn't that hard to do, and doesn't modify the original in anyway.
And I still have the manual in printed form somewhere in my archives. Maybe I should scan it, unless archive.org already has a copy ...
I know vi/vim pretty well, but I've never bothered with emacs much since the multi key stroke navigation just seems so unergonomic to me.
- https://lmno.lol/alvaro/native-emacsmacos-ux-integrations-vi...
- https://lmno.lol/alvaro/building-your-own-bookmark-launcher
- https://lmno.lol/alvaro/emacs-quick-kill-process
- https://lmno.lol/alvaro/emacs-utilities-for-your-os
- https://lmno.lol/alvaro/ob-swiftui-updates
- https://lmno.lol/alvaro/emacs-connect-my-bluetooth-speaker
- https://lmno.lol/alvaro/hey-emacs-change-the-default-macos-a...
- https://lmno.lol/alvaro/emacs-open-with-macos-app
my reasoning is that it would add another 100 years to the shelf life of emacs. its good enough i dont have any reason to switch, but if the single threaded perf i s still a bottleneck in 5 years i'd probably switch to ZED.
https://www.gnu.org/software/guile/manual/html_node/Emacs-Li...
I haven’t followed closely, but I am sure there are som valuable lessons learned there that would answer your question.