Digital PDFs

XX-7F7F1-6F

November 1979

36 pages

Original

2.6MB

Document:	TECO SIG Moby Munger 3 part2 197911
Order Number:	XX-7F7F1-6F
Revision:	000
Pages:	36
Original Filename:	http://bitsavers.org/pdf/dec/teco/TECO_SIG_Moby_Munger_3_part2_197911.pdf

OCR Text

RECE\VED,

' 1 Jl\\i ~a~()

the

COP\ES
t/1 e. TECO Spec..{al 1nte-te.~t G Jtoup
Wl

Issue # 3 - Part 2

November

1979

Contributions and correspondence should be sent to the editor:
Stanley Rabinowitz
6 Country Club Lane
Merrimack, NH 03054
or to

TECO SIG
c/o DECUS
Mail Stop MR2-3/E55
One Iron Way
Marlboro, MA 01752

Media intended for the newsletter editor should be sent to his
work address:
Stanley Rabinowitz
Mail Stop MK1-2/E~9
Digital Equipment Corporation
Continental Blvd
Merrimack, NH 03054
Camera ready copy is preferred but scribblings on the side of a
napkin will be accepted.
Ready to use material should.be prepared
.on 8 1/2 x 11 inch white unlined bond paper.
A one inch margin
~hould be maintained on both sides, on the top, and on the bottom.
Material should be reasonably clean, legible, and sufficiently dark
copy for printing.
Material intended for publication should be
mailed in a large envelope so that it can be sent without the
necessity of folding.
Readers are urged to submit articles, techniques, notes, hacks,
ideas, comments, and bug reports to the Moby Munger for
publication~
This newsletter will publish on a more.regular
basis if more readers will submit material.
It would be
especially helpful if some people would volunteer to write
certain columns regularly, such as the Beginner's column,
or. the TECO Techniques, or the.NOTES FROM XXX, etc. Volunteers
are urged to send t~eir names in to the editor.
©1979, DECUS
It is assumed that all articles submitted to the editor of this newsletter are with the authors' permission to. publish in any DECUS publication.
The articles are the responsibility of the authors and, therefore, DECUS, Digital Equipment Corporation, and the editor assume no responsihilitv nr lbhilitv for articles or information appearing in tho document.

NOTES FROM THE EDITOR
This is a continuation of the August issue. The index. for .
this issue is contaiped in that issue. :Th~_ entire ~ssue was
ready in July; however when I g.ave DECUS a 7 4.::..page iss.\.le, ,
they told me that the TECO SIG was 6nly bud~eted for t~o ;
40-page issues a year. Consequently, I decided to split :
the issue up into two parts. Since the 'y~~~· ends around June,
this means that this will be the last issue you ··see .for 6_
months unless we can get our budget changed. We'll talk about
this problem at the SIG meeting in San Diego. Part of t~e
problem is that we had no chairman at the .. timt? c;>f ~udgeting i
so DECUS assumed the number of issues we wanted this year
would be the same as the number we published the previous year.
we wouldn't be in this problem if we had a strong leadership.
Also, I was given the option of reducing the text 50% (thus
printing 'two pages on one side of a sheet). That would.have
allowed us to publish another t~o issues. However, I did not
like the looks of other newsletters that have done such a
photoreduction; so without a steering committee to discuss
such policy I made the decision on my own not to go to half
size.
If y~u want more to say on.such topics, volunteer to
be on the SIG's steering corrunittee.
1

This is the only page. in this newsletter to be typed in
November; the rest was typed in June. As of November 1, TEC0-11
and TEC0-10 are still not in the DECUS library.
I don't know
what the problem is; both have been done and working for over
5 months now.
In any case, copies will be down at the symposium
(I will hand carry them there if no one else does).
Only two solutions submitted so far for the current problem of the
Month (the macro that locates itself). Deadline is not yet up
thouqh.·
The symposium itself looks pretty good for TECO people. We have
a lot of interesting stuff. There are a few errors in the
official program; but nothing serious. The introductory TECO
tutorial will be given by Pat Hall and J. Hayes. All the TECO
tutorials t~is year are being well-prepared in advance. They
should all be good - not our usual last-minute affairs.- We were
supposed to have two EMACS tutorials. 'l'he SIG wanted this very
mu'ch - we believe EMACS to be the up-and-coming great new concept.
As of November 1, I have not heard from DECUS why these tutorials
did not make it into the program.
If possible, I would hope to
still get them scheduled (as birds of a feather session if
necessary).
The SIG business meeting will also be a TECO-roadmap
session; so be there to here about any changes in the schedule.
Because meeting rooms are suppose to be locked up at 11 PM,
DECUS would not let.me officially schedule a TECO night owls of
a feather session for midnight Thursday night; but I hope to still
hold the session either in our campground (if it's open) or in my
hotel room.
- Stan -

OVERVIEW OF EMACS

Page 42

BY RICHARD M STALLMAN

EMACS is a display editor. This means, first of all, that you see
the result of each command as soon as you have finished typing it. Ue
try to make this true of th~ smallest possible units of typing. So.
for example, there is no "insert" command:
to insert a charac'ter. you
just type it, if it is a printing character (there is a command to
quote other characters). As soon as you have ·typed it, it appears on
the screen in the right·place. If you make a mistake, just type
Rubout, and the character before the cursor will disappear. Other
commands wi I I move backward or forward over characters, up or down
I ines or to the beginning or end of the current" I ine, to the beginning
or end of the buffer, etc.
This much applies to al I display edi'tors.

Any display editor at all provides a tremed~us advantage to the user
over printing-terminal edito~s. Most users 6f display editors (EMACS,
.TVEOIT, Stanford ETV) find that they no longer need to make paper
I istings.
Also, it becomes very easy t_o cc>mpose new programs on line.
However, EMACS offers the additional feature of self-documentation.
Not only are there the obvious commands which ask what a given _
character does; there is the comman~ "Apropos" which can tel I you
about al I commands ~hich have a given string in their names.
If you
have forgotten how to read in a file, just do ~X.AproposoFile. This
11
\.Jill tell you about all commands whose names contain 11 File , one of
~hi ch wi If be the one you seek.
Additionally, there is an EMACS
subsystem which is used for perusing tree-structured documentation
f i 1es.
It is ca 11 ed INFO; one 1-1ay of entering it is to type ?" &.1hen
EMACS starts up.
INFO offers a tutorial on how to use INFO. as J.Jel I ·
as documentation on EMACS and whatever else you care to add to the
tree.
What distinguishes EMACS from all other display editors ia that it
is conveniently programmable:
it contains an interpreter for a
pow~rful programming language (called TECO, but much more suitable for
system programming than DEC TECO~ See the appendix). T~is makes it
very easy for new commands to be written. An individual user can add
a few commands of his own. or an ambitious person can ~rite a sharable
I i brary of use fu I commands uh ich everyon·e can use. True, s i nee any
editor is a program. if you ha~e the source you can modify it. but in
practice a much greater sophistication is possible in EMACS programs
· than in assembler-language modifications to TECO.
0

For example, you can program EMACS to understand the syntaw of your
favorite programming language. At MIT we have found useful commands
for editing Assembler language, text (and input to text justifiers),
L1sp, TECO code (for EMACS), and PL/l. The user can select a •major
mode" such as "Text Mode", "Lisp Mode", etc., and each mode redefines
a fe&.J commands to be appropriate to the language (a user or alte .can
add ~ew major modes). In addition, there are several comment
·
manipul~tion commands which can be used for any language, because the
major mode tel Is them wbat strings are used to begin and end comments
in the language you are using.
In ".Ml DAS Mode" (our assemb I er is ca II ed MIDAS), there are commands

for moving to the next or the previous label, and for moving ~o the AC
field or address field of an instruction.

For Lisp, we have commands

~hich indent code according to its.nesting, ~~ich move over or up and

down in list structure, which go to the beginning or end of the
function you are working on, etc. For Pl/l, there la a very hairy
indentation program which figures what sort of statement or compound
statement you are in the middle of and indents the line appropriately.
The techniques used for PL/l can easily be adapted to most other
algal-like languages.

Page: 43

For editing text. ~e have commands that understand sentences.
paragraphs, and ~ords, as ~el' as commands to tilt and justify text.
Also. there is "Auto Fill" mode. in l..Jhict) tines are automatically
broken when you type space, to fit a specified margin. I .am using
auto f i I I mode now. For text justifiers, ue have commande to insert
and delete the text justifi~r commands· that specify font selection and
underlining.
·

To save the user from having to select the right major Mode by hand,
each file can specify the major mode to be used to edit It. The
string "-*-Text-*-:-" at the front of this file has that purpose. It Is
also possible to specify in a file the value of any of t~e many
options and parameters that control EMACS. The values thus specified
wi I I apply whenever EMACS is looking at that file. Als~, a user can .
specify in his start-up file that certain special optlons_shoufd be In
effect whenever he (and only he) is in a given major mode. For
example. my start-up says that auto f.if 1 should be on whenever I_am in
text mode. 1 f 1 read in this file, auto fi I 1 is turned on, but tf you
read it in, you ~on't be in.auto fill unless you ask for it.
You can edit tuo windows at once. The screen ls split across the
middle, and the top and bottom can show two differ~nt files or t~o
parts of the same file. You might want to p~t a I 1st of tas~s. or a
.FlLCOM, in one part of the screen, while making the chang~s tn the
other ha If.
With large systems split up into many files it can be hard to
remember what ts in each file. With the EMACS TAGS package you don't
have to.
It remembers al I the location of each function def ini ti on or
label in one or several files, and ~ill jump directly to the
definition of the function whose name you specify, s~itching files as
necessary. TAGS uses a special file called a "tag table", which y6u
must construct by cal ling a program provided for that purpose. Since
it contains the character-position in the file of each tag, tags can
be found without searching the whole file. The tag table is not
updated automatically when the file is edited, but this is not a
problem. because a small inaccuracy mandates searching only a small
part of the file, not the ~hole thing.
One of the most avid EMACS users at MIT has written a "word
abbreviation" package~ If I defi~e "pk" to be an abbreviation for
"package", then I can just type "pk" and it will expand automatically.
If I' type 0 PK", it u i I I expand to "PACKAGE". "Pk 11 1-1 i 11 e><pand to
"Package". Some people have large files of abbreviation def_initions
~hlch t~ey always use.
Word abbreviation mode is one example of a sharable EMACS library
by a user. Several users at MIT have uritten libraries for
general use. including the PL/l feature described above. Every
I ibrary contains documentation for each command in it, so that the
self-documenting features continue to uork.

~r1tten

EMACS is available to anyone for use on Twenex, on a communal
sharing basis. Any generally useful improvements that you make, you
must give back to MIT to share with everyone else, just as we have
shared with you. If you plan to make improvements, you.should get in
touch uith MlT and other sites so that you can find out ho~ to make
them as widely useful as possible.· Of course, uhoever Is actually
Mi I I ing to do the work has more say than everyone elae~ but it is his
responsibi I ity to consider ~he needs of all usera.

If you woul~ l~ke to talk ~ith other T~enex users of EMACS, you can
contact the Computer Corporation of America. 575 Tech Square,
Cambridge, MA 02139, or Geoff Goodfelto~ ai the Stanford Research
Institute. There are also users at DEC (I hear that the Maintainer of
DEC POP-10 TECO is one of them).

Appendi><: TECO.
.

Some of ~tau may have tried programming in DEC TECO.

lt is very

tedious, since you can't have very many variables or very· many
subroutines, and al I error conditions must be detected explicitly.
MIT TECO provides arbitrarily long names for variables and
subroutines, thus al lowing reasonable programming practices to be
fol iol-led.
In add. ~ion, errors can be handled in several ways.

First

of al I, you can put an error-catcher around some commands;

an error
inside ~i I I return to the ciatcher, also returning a pointer to a
string containing the erro~ message. Second of all, the user can
supply an error handler which will be called in the environment of the
error.
In EMACS, this error handler allo~s you to look at the buffer
and at the stac~ of macro calls. Finally, if you don't do anything,
the local binding stack (known 1n former times as the q-register pdl)
~i 11 he unwound, so that global values of variables are never loat.
Wi tho~t these error handling features, it ~ould be impossible to ~rite
reliable system programs.
·

Also, data structures have been made more flexible and efficient.
I t ·may work even in DEC TECO to do 01U2 to rnove a string from 01 to
02. This is because strings are represented by pointers. In MIT TECO
we have taken a~vantage of this to pass pointers to strings much of
the time.
In addition, yau·can have pointers to other sorts of
objects: buffers, for eKample (there can be any n~mber of them), or
vectors of other objects (good for symbol tables, etc). Every buff~r
has a moveab)e "gap" which is invisible to the user; this ia used to
/nake insert·ion and deletion in a local region of the buffer much
faster.
Insertion just uses up gap, ~hi le deletion increases the gap.
Of course, you must move the gap to ~here you ~ant to insert or delete,
but that's better than moving half the buffer for each insertion or
delatio~.
The abi tity to Lse a temporary buff~r for intermediate
results reduces core shuffl i.ng eveh more. As a result. when you edit
a small part of a large file, only the part you are editing needs tp
be swapped in.
Finally, MIT TECO contains an integral display processor designed to
display an up-to-date picture of the buffer, using as fe~ characters
as possible to update on whatever sort of terminal you have.
The best testimony to the effectiveness of MIT TECO as a system
programming language is that about 8,000 tines of code have been
~ritten and debugged using it.
Historical note:

PDP-10 TECO {actually, PDP-6 TECO} was originally written at the MIT
Artificial Intelligence Laboratory in 1964 by Greenblatt, Nelson and
Holloway. Some time later, DEC took a copy, but development of TECO
at MlT did not cease. In 1973, 1 added the real-time display
proc·· ~c:or, and the abi Ii ty for each character to run an arbitrary
program.
Immediately, many users began to write packages of such
programs:

TECMAC, RMOOE, DOC, MACROS, THACS, Russ-mode,.~.

This

explosion of TECO program development allowed many ideas for useful
commands and comm~nd langu~ges to be tested. Until 1976, I played no
part in this work, being content to add to TECO the features requested
by the maintainers of the ~ubsystems. EMACS ~as b~gun in 1976,
drawing on what had been learned from all the other TECO program
packages, as wel I as borro~ing the code for some complicated commands.
Some of the older programs have been supplanted entirely; aome have
been re-implemented as cu~tomizations of EMACS to save their old users
frnm h~vinn t~ re-learn the command set: RMODE and DOC are

Page 44

Page 45

RANDOM IDEAS

by Stanley Rabinowitz

The following are my (subjective} ideas for possible future
enhancements to TECO ('I1ECO-ll in partic"ular) . Comments,
improvements, and other suggestions are welcome.
l .·
2.

m, n "E could mean execute the following commands if m=n.
m,n"N could mean execute the following conunands if m~n.

"X could mean execute the following commands if a numeric
argument exists.

n"#~ or perhaps n:"~
(where X represents some conditional)
could mean execute the following commands if n is NOT
satisfied by condition X. For example, n"#D would mean
execute the·following commands if n is not the ASCII code
for a digit.

Multiple I/O channels could be implemented in a more
consistent and mnemonic manner than the current scheme
as follows:
nER$ could mean switch to input channel n,
and nEW$ could mean switch to output channel n. Thereafter,
ER's and EW's without a numeric argument would work on
the current channel.
In addition, nERfile$ and nEWfile$
would open a file for input (or output) on the specified
channel.
nt.Atext@ ·w~uld read input from the terminal allowing

7..

1¢.

normal editing facilities.
"text" would be used as a
prompt.
Input is inserted into the text buffer before
dot and is terminated when the character whpse ASCII code
is n is entered.
If n=O, then input is terminated by
crlf.
If n=-1, then input is.terminated by$$. Until
the input was terminated, normal editing features, such as
rubout and CTRL/U would be permit~ed.
nEC could tell TECO to shrink or expand to nK by"t!·es
(words?) of memory. ~EC would mean to shrink to the
smallest possible size.
The match control construct CTRL/E M could mean search
for multiple occurrences of the next match control
construct (as in TEC0-1~) .
:Otag$ could be a subroutine call. Control would transfer
to the specified tag and upon execution of some as-yet
undetermined command which would correspond to a RETURN
conunand, control.would resume following the :Otag$ command.
Such a construct would greatly enhance the easibility of
writing good structured TECO code. The command to return
from subroutine might be something like :0$ although,
technically, this should mean call the subroutine whose
name is the null string (represented by a ! ! tag ) . ·
Frequently, at the beginning of a cnmplicated macro, ·it
is necessary to push away a lot of Q-registers and at the
end of the macro, to pop them back. Such situations could
be simplified by commands like [<A,B,C,D, ... > and
]<A,B,C,D, ... > which would push and pop a set of Q-registers,
respectively.

Page 46

In the last issue of ·the Moby Munger, the editor presented
~n editorial chiding the differences· between TV and TECO.
,...-ie following article is a rebuttal by Eric Osman, whQ
reviews some of the features of TV and argues that the merits
of TV override its incompatibility with TECO.
Further
comments, both pro and con are welcome.

ABOUT TV

by Eric Osman

In the March 1979 issue of the MOBY MUNGER, some menti~n was
made to TV and its relation to TECOc As.the father of TV, I
would like to shed some further light on TV~
TV is a TECO that runs on TOPS2~e
It features a window
feature that shows the current buffer portion beinQ edited,
with a cursor to indicatP. where the pointer is.
Tnts is
particularly useful on video terminals, on wnich TV knows
how to update only the parts of the window that ·have
changed, when a complete repaint of the screen can be
.. avoided. Note, however, that the window feature ls ·also
available on hardcopy terminals.
I arrived at DEC f iVe years ago and immediately was
disappointed that no video TECO existed
I proceeded to.
create TV from VTECOc
VTECO !s the BBN TECO that was
~unning on TOPS20 wnen I arrived at DEC.
0

~re

some ot the major points about TV:

ar~

BACKUP SYS1'Ei.;M:

Every fifty keystrokes, every character the user has
typed ls appended to a file called COMMANDS.IV.
If
the system crashes and is restarted, the TV user can
recover nis work, losing at most fifty characters,
merely by 1ssue1ng a command telling TV to read the
contents of the tile as typein.
This method has the followlnq advantages
others:.
o

Standard

some

has no automatic
backup.
tne system crashes, you ·lose lot.s
of work, unless you continually remember to
"save your file" per1octica11y. TV lets you
forget about such worries and just type!
Hence

over

T~CO

Some editors feature an auto-save feature,
which saves a copy of the entire tile
periodically. This has several pro~lems.
First of all, if it's a large tile, you may
not have disk space for the extr~ copies.
Also, . saving
an
entire co PY is very
expens1Vee
TV's savinq
ot
only
your
keystrokes
d11rinq
thE~
current ed.it1ng
session uses much less disk space
and

effort.

Pag(~

Another method

buffer to a file:•

bacl<up

is · to

map

the

This has the Problem that

after <?- cr~sh, you nave to figure out
exactly when in relation to your tvptnq, the
system

crashed,

you

can

know

Which

keystrokes "m3de it".
FORGETTING TO TYPE I:
All of us TECO users have .at some point forgotten .to
type I at the beginning of an iniertion string.
After hittina two escaoes at the end ot the command
string and getting an.error, TECO lets you recover
if you type asterisk followed
by
a
letter.
Unfortunately, .it vou typed one non-aster!sk,you
ar~n·t Qiven the option of deleting
it and typing
the asterisk.
TV has the ;G (semicolon G) command
which insert~ the last command string into the
buffer that was lonqer than fifteen characters.
Hence short command strinqs or extra keystrokes
follow1nq a long insertion that failed don't prevent
you from-recovering your typing~
·

LARGE BUFFER AND HOLE:
Instead of the standard TECO method of edit1n9 page
page, TV works most conveniently if you issue the
1Y command which yanks an entire
t11e into the
buffer, edit it, and tnen use ;x to exit.

This avoids several pitfalls of TECO. for instance,
an inadvertent Y command will give you an error,
since no file is open1 In many TECOs, 1nadvertent Y
commands causP. the current page to be erased and
replaced by another page! Another Pitfall avoided
is that of misspelling a search string. In TECO, if
vou say Nrec1ev$, TECO shuffles through all the
paqes of the file, then gives an error, since you
meant NreceivS. Unfortunately, as min~mal recovery
effort, you must reopen the file. tn TV, you would
merely reissue the failing search, since no file ls
open, so the N would not clear the buffer or read in
more data.
Most TECOs allow the entire file to be edited in the
buffer at one tfme. However, this is grossly slow,
sine~
every insertion or deletion requires the
"rest" of the buffer to be slid over to make room
for the insertion, or close the qaP after
a
deletion.
TV
solves this problem with a hole, or
bubble, in the buffer. This bubble floats to the
location of an insertion or deletion. A deletion
causes the deleted characters to be absorbed by the
bubble.
An in~ertion causes the bubble to shrink.
Hence the amount of data necessary to be shuffled is
reduced to the distance between one insertion or
deletion to the next. For instance, the followinq
command strinq, which inserts a carriage return
after every sentence, could easily run a hundred

Page 48

times faster in TV than TECO:

This is because @Very time TECO inserts the next
crlf, the entife buffer to the right of where the
crlf is being put must be moved riqht two places.
In TV, only the sentence and its period need be
moved.
INCOMPATIBLITIES WITH TECO:.
TV 1~· quite similar to TECOr and anyone starting to
use TOPS20 who ls familiar with TECO, can learn TV
by merely knowing a short list of differences.

Most of the 1ncomPat!bilit1es are due to the
that Tv· was created from VTECO, which ls BBN's
ant! already had the incompatibiliti_es.

fact
T~co,

Althq~qh

having a compatible TECO is desirable, a
number of the differences, I believe, are better in
TV ·than TECO.
For instance, having the
file
commands

and

;,w etc.

instead of ER and Etti

reduces the chances of runaway insertions causinq
irreparable - damage.
CA runaway insertion is one
~issing the I before it).
·Another incompatibility
1s
that
searches , in
iterations are not assumed to be "colon~ searches in
TV, as they are 1n TECO. This is nice because in
TECO if_ you forget a semicolon after your search
command in ari iteration, you get into an infinite
loop. In TV, you'll Qet an error.

Other
1neompatib111t1es
resolved.

could·

conceivably

All new commands that I ·have added to TV, I've tried
to be compatible with TECO. For instance, TV now
has most of the control•E search features.
The original BBN VTECO, from which TV was created,
started command execution when one escape was typ~d
instead of the two that TECO waits tor. I modified
TV to wait for two~ as TECO does. Th~s was done ln
response to many people complaining that
they
accidentally typed escape in the middle of a command
~trlng (such
as in ffn iteration) and execution
started prematurely. tn BBN VTECO, you type CTRL/D,
which
~choes
as
s,
when
you
need
a
non•command•termtnatinq~escape.

Page 49

FUTURE OF TV:

Although TV shows the edit1n~ buffer on the video
screen, TV can not be used as a true video ~d1tor
very conveni@ntly. The true video editor allows
keystrokes to cause immediate effect on the screen.
For instance, if the pointer is in the middle of a
line, and you type x to the true video editor, the x
immediately -appears in the line.
Likewise,
a
subsequent DELETE character would eause the X to
disappear, and the rest Of the line would instantly
slide left to f 111 the gap.
True video editors also allow the user to associate
keystrokes with macros. For instance, you may want
to be able to tYpe CTRL/S to save the current
Then you move the Pointer to a new
sentence.
saved
location and type CTRL/V to insert the
only
one
of
a~
1nf
1n1te
here.
This
is
sentence
number of examples.

A qood example of a video
.at MIT.

~ditor

is EMACS

developed

Either TV will grow the ability to do what eMACS
·does, or .·EMACS Will become the popular editor. It
is not c1ear yet which of these will happen.
However, _true video editors are definitely the
easiest editors to usee

DIRECT.TEC - Installation Procedure

(continued from page

2~)

To invoke DIRECT.TEC, you put the following equate in your
login conunand file:
DIR:=='MUNG'

[loc]DIRECT,

where [loc] is the name of the directory in which DIRECT.TEC
resides. Note that the conunand ends with a conuna and must
occur after the definition of MUNG.
To use DIRECT, you would issue a command of the form
DIR f ilespec
where "filespec" may contain wildcard *'s, ?'s, or %'s.
For example:· DIR A*B.FOR or DIR [*]·FOO??X.*
"?" and % are equivalent.
11

Page 50

Dear Stan:

April 11, 1979

In the most recent Moby Munger ( #2, March 79) you published a commen.tary
from Tom Mcintyre on the advantages and disadvantages of KEO and TECO.
I think Tom did a good job of comparing the two from the editor
.
designer's point of view, but he did not say much about some of the issues
that I find most compelling when it comes down to· picking an editor.
The most sighificant of these issues for me are:
1. TECO exists (as you have extensively 'documented) in compatible versions

on all the DEC mainframes and operating systems. It seems very important here to be able to move from RT-11 to RSX-llM to RSTS to OS/8
·to OS/78 without having to learn several different editors. Only masochists and gurus do that. It is particularly difficult to get
·
compatible editors when going from -8 ba·sed systems to -11 based
systems. Compatible TECOs are either distributed free with the systems
or are available from DECUS at low cost for all these systems. On the
other hand, KEO is not. ·It may or may not be available for RT-11.
I have had trouble finding out about cost, etc. Somewhere down the road
it might be announced for the other 16/32 bit systems although I know
of no DEC commitment. I expect it to be at least a few years before
it ts available on all the -11 systems we use. Even then we still have
the issue . of cost. As far as I know, there is no expectation of getting
KED for the 12 bit systems. As noted above, this is a major disadvantage
for me.
2. TECO rµns on any terminal from a teletype to a VT-11 interactive graphics
display system. KEO only works on a VT-100 (it might get VT-52 support but don't bet too much on it). We have a teletype, DECwriters,
a VT-52 and a Vf-11 that need to be used with the editor we pick.
Exactly because, as Tom points 6ut~ KEO is a DEC supported product, it
is not readily available to the users to modify or to add support for
other devices. The decisions on which machines, operating systems and
terminals wi·ll be funded for support, and when, is driven by DEC
marketing considerations rather than existing users needs.
3.. KED would be, most certainly, more user oriented, more 11 learnable 11 , safer
<md more satisfactory for the vast majority of users than TECO is. The
key to this quandary for us is the TECO ·video terminal support that is
now available on all the systems we are interested in. With it, a
version of the VTEDIT macro can be used on all systems with a VT-52, ·
VT-100 6r VT-11 giving a screen oriented editor that can be learned
by naive users in minutes. The new user will be able to move across
all our systems with almost no further training.
1

When KED is available on at least all the -11 systems we use, it is low
cost, and works with the terminals we have, I will give it very serious
consideration as an alternative to TECO. When KED is supported in this
way on our 12 bit systems (i.e., pre Omnibus included) I will almost cer~ ~inly.change over to it.
Until then, TECO continues as the only useful,
'-rtable editor for me and my user community.

Very t~ul~yurs,

~!o~~···

Page 51

MORE ERRORS IN TECO REFERENCE CARD
Since the last issue of the Moby Munger, we have noticed a
few additional errors in the TECO Pocket Guide (AD-D530A-TK).
These are described below:
5.

Page 3, the explanation of the EC command, "moves the
remainder of the input file to the output file ... " should
read "moves the text buff er and the remainder of the input
file to the output file ... ".

Page 4, the 1£st of legal switches urider TOPS-1~ is
missing some switches.
In addition to /GENLSN and /SUPLSN,
TECO also accepts the switches /SCAN, /NOCREATE, /NOWRITE,
/SYS, and /NEW.

Page 11, in the explanation of the ES corrunand, "an effective
mV·conunand is executed when a search is completed" should
read '-'an effective mV conunand is executed when a successful
search at top level is completed".

8 . . Page 12, under CASE CONTROL, we are missing the ¢tv and ~tw
conunands. These should be printed in red since they are
implemented in DEC's TOPS-1~ TECO only. These commands
are identical and take TECO out of case conversion mode.
9.

Page 13, the search string build conunands CTRL/V and CTRL/W
should both have a note that reads "use of these constructs
in a search string causes the search to be executed in exact
case mode under TOPS-1~".

l~.

Page 14, the explanation of the ?MEU error message should
read "Macro ending with t", not·"Macro ending with U".

11. Page 16, the graphic for the ASCII character whose octal
ASCII code is 14~ should be~, not~
(The accent is
slanted the wrong way.)

ERRORS IN MOBY MUNGER
The following errors in past issues of the Moby Munger have
been noticed:
o

In the NOTES FROM RT-11 on page 6 of issue #1, we said that
under TEC0-11 V28, the ntW conunand (with ns-2) performs a
screen update. This is wrong.
It merely tells TECO that
lnl-1 lines have changed on the screen. The window is not
actually updated until the next -ltW conunand (notwhen TECO
goes back to prompt level). Note that in versions of TEC0-11
after v28, the window conunand is W rather than tW.

DECUS has informed us that the program, DECUS-L-125 described
on page 53 of issue #2 is no longer available from the DECUS
program library.

Issue #2, page 7, "TECO.TEC" should read "WILD.TEC".

Page 52

VERSION 1(165)

WPI TECO

WPI TECO is a modified version of DEC's TOPS-1~ TECO v23.
t was modified by Al Johannesen at WACCC who -is also the
~urrent maintainer. ·should anyone desire a copy, Al will
send a copy to anyon~ who supplies the media.
His address is
Al Johannesen
Worcester Area College Computation Center
Worcester Polytechnic Institute
Worcester, MA 01609
We described.many of the features of WPI TECO in the last issue
of the Moby Munger.
The following new features brings us up-to~ate.
This information describes WPI TECO vl(l65) and is
current as· of June 1979.
·
13.

lEP sets the page number display fiag.

14.

64EP allows an EB command to be prematurely terminated
(current output file remains as a .TMP file).

15.

128EP re-enables normal use of the Y command.

1.6.

EY is the same as Y but must be used instead of Y at command
level (unless 128EP is set) .

17.

ntP pages out to page n.
Truth-in-paging mode (4EP) must be
set.
~tP pages to the last existing page in the file.
Backward paging. ~s not permitted ..

18.

:=and :==were implemented.as in OS/8 TECO and TEC0-11.
·(They suppress the crlf) .
: Xq was implemented as in
TEC0-11.
(Appends to Q-register.)
:T and :EQ are like T
and EQ but control characters (including carriage return,
liri~ feed, and tab) will be typed out in up-arrow mode
regardless of the prevailing typeout mode.

19.

<CTRL/G>.::::FORM-FEED> is an immediate mode command that retypes· the entire command string.
(<~TRL/G>* on some other
TECOs.)

2%.

nV is identical to -nT(n+l)T .

......

·~

Vis· the same as grT.

(I prefer V to be the same as lV and to define nV as (1-n)TnT.)

21.

nFDtext$ (Find and Delete) is equivalent to nFStext$$ .
m,nFDtext$ is equivalent to m,nFStext$$ .
·
(Andy Nourse-'s TEC0-10 has implemented t:his version of the nFD cormnand.)

22.

To save space, tF, tH, and EM were removed.

23.

When using : or@ modifiers, any numeric arguments must go
after
the : or @. · (This seems lJ~ke an unnecessary restriction.)

24.

An additional Q-register, #, has been added.
This register normally contains the filename last referenced by
an ER or EB command.
(This is a true Q-register, unlike the
pseudo-Q-register, *, used by TEC0-11 for a similar purpose.
I think
that real Q-registers are better than pseudo-Q-registers.
* is a·pseudoQ-register under TEC0-11 because it can only be used in limited contexts.)

25.

The source is now written in FAIL. Also many internal
speed-ups have been made.
(Some speed-ups, like use of AC
loops are of use on~y on KAl~s.)

Page 53

XTEC
XTEC is a TECO compiler for the DECSYSTEM-1~.
It runs under
TOPS-lS or ITS (under the compatibility package).
This program
is available from the DECUS program library~
Its order number
is DECUS-lS-264.
It was written by Jack W. Krupansky and
Mark R. Crispin at Stevens Institute of Technology.
The version
in the DECUS library is %~(427), submitted in February of 1976.
As far as I know, this is the latest version.
XTEC is patterned after Stevens TECO (version 123).
It is
a compiler rather than an interpreter· (as most other TECOs are).
When you start execution of a command string, or when you execute a Q-Register, XTEC compiles the TECO code into PDP-1~
machine language.
Re-execution of the same macro does not cause
recompilation unless the contents of the Q-Register has changed.
XTEC execution is over an order of magnitude faster than Stevens
TECO execution· of the same TECO program.
Stevens TECO (Vl23) was described on pages 33 and 34 of the
last issu~ of the Moby Munger.
The differences between XTEC and
Stevens. TEco·are described below:
1 ..

A value cannot be returned from a macro.
If you colon the
macro, both the first instruction and the entire macro will
be "coloned", returning a value of -1 if it succeeds, ~if
it fails.

Values· passed to a macro must be consistent in number in all
usages of that1nacro.

XTEc·supports Q-Register names of up to six characters.
If
more than one character ~s used, then the name must be
enclosed by parentheses.
For example, the command M(PRINT)
executes the macro whose name is "PRINT". Allowing symbolic
names for TECO "variables" makes for much more readable code.

The EI command uses a temporary Q-Register and does not smash

If a compile-time error is detected, the conunand is aborted
without doing anything.
Therefore the conunand string HK=$$
will generate an error message and will not destroy the text
buffer.

XTEC has different f ilespec blocks for each I/O command
(used for filespec defaulting).

The W command of Stevens TECO is not in XTEC.

. 8.

The EO value of XTEC is ~ (.functionally equivalent to EO
level 2
Stevens TECO) .

The Superseding Existing File message has been changed to use
the error segment.
To allow usage of "/" and "?", X'l EC will
type a fake prompt and wait for a continuation character to
continue execution of the XTEC command.
This character is
ignored.
(This is very confusing to the user.)

1,0.

XTEC supports the following switches in filespecs: /SIXBIT, /ASCII,
/OCTAL, and /PROTECTION:nnn.
/SIXBIT reads the file in pure
SIXBIT.

Page 54·

HARVARD UNIX TECO .(HRSTS TECO)

"9llarvard UNIX 'rECO runs under the UNIX opera ting syst~m on a
PDP-11.
It was written by Bob Case, Peter Langston, Bruce
Borden, Brent Byer, and Tucker Taft using the Harvard-Radcliffe
Student Timesharing System (HRSTS). A manual, written by
D~niel Waugh and Ken Stiong, is available from the HRSTS
System Manager.
For more information, or to obtai~ a copy of
·this TECO , write to
Stephen Dyer
HRSTS System Manager
Harvard Science Center
1 Oxford St.
Cambridge, MA 02138

ll~a

Educational institutions can obtain this TECO for a nominal fee.
Harvard UNIX TECO is based on a 1973 version of MIT TECO and is
written in MACR0-11.
There is a possibility that this TECO will
be expanded and recoded into C in ·the near future.
The following list summarizes the advanced ·features of Harvard
UNIX TECO by specifying those features that are not in other
traditional TECOs such as DEC's TOPS-1~ TECO.
1.

,. .
\

LF, if typed as the first keystroke after TECO's prompt,·
performs an effective LT command.

$ or·+, if typed as the first keystroke after TECO's
prompt, performs an effective -LT command.

CTRL/A immediate mode command is equivalent to RUBOUT.

The immediate mode command CTRL/B aborts the command string.

Space eats numeric arguments.

6.
7.
8.

EDfile$ deletes the

EQ (Quit) and tDtD cause an immediate exit from TECO.

l~.

nV is equivalent to -nT(n+l)T .

11.

:T and :V are like T and V except that the type-out.
shows non-printing characters.
For example, space prints
as <SP> and the character whose ASCII code is 1 (CTRL/A)
prints as <SOH> .

12.

The K command riot only kills text but in addition.performs
an effective : X@ conunand.
'1 his puts the killed text into
pseudo-0-Register @.
Thus, you can get the deleted text
back with a G@ command.

sp~cified

file.

EY replaces the traditional Y (Yank) command.
ET (Top) repositions the input file to the beginning of
the file.
V is the same as

3 ... E+text$ replaces the traditional +text$ destructive global
search command . .

14.

nFDtext$ finds and deletes the nth copy of the specified
text.

rage 55

HARVARD UNIX TECO (coNr.)
15.

Match control constructs have been shuffled around a bit.
CTRL/Q matches any character, CTRL/P matches any separator,
and CTRL/E is the quote character.

16.

Upper and lower case Q-Register names are distinct.
(This has proven to be a lose in other TECOs.)

17.

:Xq is like Xq but the extracted text is deleted from the
text buffer.

18.

The Edit Macro .command, EMfile$q$, reads the specified file
into the specified Q-register.

19.

The Exchange Environment command, EEq, 'moves' you into
the specified Q-Register.
That is, the text in that QRegister appears to be the normal editing text buffer and
the original text buffer has been saved away in a pseudo.Q-Register. While in this mode, TECO's prompt changes to
q* where 'q' is the name of the Q-Register you are editing
in.
The EE$ conunand restores the context to the main
·
text buffer.
(A very useful command.)

·2¢.

+ returns the negative of the number of characters in the
last search string.

21.

Q+ returns the position of the beginning of the last match.

22.

Q] returns the position of the pointer before the last search.

23.
25.

Q[ returns the position of the pointer before the last insert.
tE flag is a 1 if end-of-file has been reached.
twtext1-W types out the· specified text (traditional. tAtexttA).

26.

tL clears the screen on a video terminal.

27.

E!$ (E-bang!) spawns a sub-shell·. When the user logs out
of the sub-shell, control returns to TECO.

28.

E!cmd$ spawns a shell to execute the single command specified and then control returns to TECO.

29.

All flags are handled in a consistent manner by the commands
ESf, ECf, and EVf where f is the flag name.
(Each flag name
consists of a single character.)
ESf sets the specified
flag, ECf clears the flag, and.EV£ returns the value of the
flag. (¢ represents off and -1 represents on.)
A description of a few of the flags follows:

24.

Flag

Description

controls creation of .BAK files

controls whether or not typing CR on terminal
enters CR/LF or LF only

form feed flag

long error message flag

trace mode (causes user commands to be logged in a
file called TECO.TRl\CE)

Page 56

rv is a TECO variant that runs under the TOPS-2~ operating

,ystem on a DECSYSTEM-2~.
It has many features in common with
other TECOs available at DEC, but it differs from 'traditional'
TECOs more than the other TECOs that have been described in
·recent issues of the Moby Munger. The main problem is that
the comm'-rn commands R, ; , E, and F behave radically different
under TV than they do under traditional TECO.
The information I have on TV is current as of July 1978.
TV contains the following features that ·are not in other DEC
TECOs:

~he

;Y means yank the entire input file into memory.

@J moves the text buffer pointer to the center of the screen.
(TV sort of maintains a window into the text buffer on the
screen of a video terminal.)

;U ~eans finish edit but stay in TECO (equivalent to the EC
command of.TEC0-8 and.TEC0-11).

Negative arguments to a search cause the search to proceed
backward. The pointer is left positioned at the beginning
of the string found.

;G gets the last command string that was longer than 15
in length and inserts it into the text buffer.

charac~ers

6. · ; Ef ile$ ·causes subsequent input to TECO to come from the
specified file.

'•1. nRa$b$
causes the next n occurrences of string "a" to be
replaced by the string
"b"~

TOPS-1~

1·G means exit from TECO (as in DEC's

TECO, EO level 1).

CTRL/G run-time comrriand aborts execution of the currently
executing TECO command. (Why isn't t:his CTRL/C?)

l~.

Backspace, if typed as
the first keystroke after TV's
prompt, performs an effective -LT command.

11. Linefeed, if typed a.s the first keystroke after TV's prompt,
performs an effective LT command.
(The above two immediate mode commands are now in TEC0-11 V33 and
TEC0-8 V7.)

12. tL clears the screen.
13. Carriage return eats up any pending numeric value.
14~

CTRL/Q is an immediate mode quote character.

15. CTRL/U if typed as the first character in a line of a command
string, deletes the previous line of that command string.
16. $ (dollar sign) clears the current numeric value.
17.· ;D puts a heading at the beginning of the buffer and then
issues the ;_U command.
~8.

; H exits from TECO (Why isn't this tc?) .

TV (CONT.)
19. ;Mqfile$ loads the specified file into Q-Register q.
2~.

;P combines the .0A command followed by a C (returning the
value picked up) .

21.

;S saves entire buffer then closes file.
m,n;S saves the
portion of the text buffer between pointer positions m and n
and then closes the output file.

22. Qq;T types the

~tring

contents of Q-Register q.

23. ;Tmessage$ types the specified message.
24. nE is the same as nL2R, that is, it is like L but is used to
position yourself at the end of a line (before the CR/LF)
rather than at the beginning of a line.

25. n:K is the same as nK but doesn't kill the final CR/LF.
26.

:L mov~s the text buffer pointer to the end of the current
line (before the terminating CR/LF) .

27. Failing searches don't move the text buffer

p6int~r.

28. V, nV '· and m, nV are the .same as the corresponding T commands,
except that the specified range of text is difPlayed on the
screen.
·29. The hXq and m,nXq commands delete the extracted text from the
text buffer automatically.
3,0. Searches in iterations are not automatically colon modified.
31. A period immediately following a digit string acts as an
'octal point', that is, it causes the number to be interpreted
in base 8.
32. There is a large number of W commands, of the form Wword$
where 'word' is mnemonic for some TV feature.· For example,
WVT,05$ telis TV that you are editing on a VT.05 terminal.
WFLAGLOWER$ institutes lower case flagging.
(,0EU on other
TECOs.)
The following TV commands are exactly equivalent to (or very
similar to) the corresponding TECO commands:
TV command

TECO command

TV command

TECO command

Ra$b$

FSa$b$

tAtext$

tAtext@

;Ftext$

+text$

n;N

;Rfile$

\
ERf ile$

;Ttext$

tAtext@

;Wfile$

EWf ile$

;<space>

Ft ext$

Ntext$

CTRL/R

CTRL/G<space>

One interesting difference between the ;X command and the EX
command is that ;X will prompt you for an output file name if an
output file·is not open.

Page SS

:1r TECO. <1rs rEco)
This version of TECO is incredible!
It runs on a DEC-SYSTEM-1~
under the ITS operating system at MIT. MIT is where TECO began;
t:-hey have kept way ahead of all other rl'ECOs.
For most other TECOs described in the· Moby Munger, ~. complain
that the author's have made incompatible changes with DEC TECOs.
In this case it is just the opposite - MIT TECO existed long
before DEC's versions and DEC made many incompatible changes.
Our only excuse is that we did not know what MIT was doing. ·
At MIT, TECO has grown very powerful; it is an order of magnitude
more co.mplicated than DEC's TECOs and ha~ an equal abundance of
functionality.
It is even more cryptic and packs more power per
character.
My only comments are: fantastic!
Stupendous! and
My God!
The only documentation I have on MIT TECO is an 80-page list of
commands.
If there is a printed manual, I would appreciate
receiving a copy. The information I have is accurate as of
2-Nov-1978 and describes TECO version 7~9.

There.is too much functionality in this version of TECO for me
to possibly describe .it all in this column.
Instead, I will
nly describe a few of the features (in little detail).
Future
"'I'ssues of the Moby Munger will describe additional commands.
Inclusion of a description of a command in this column in no
way is an endorsement of the merits of that command.
~ame

nt@ is the

tA, is an inclusive-or arithmetic operator.

tofilespec$ bigprints the specified filename on
open for output .

.5.

to in a search string is a special character signifying "OR",
that is, it divides the search string into two strings
either of which will satisfy the search.
·

{·P is a sort command.
The description of the arguments and
how this command works fills up an entire page and is too
lengthy to reproduce here.
Perhaps someone would like to
contribute an article .entitled "SOHTING WITH MIT TECO".

ntz inserts n random·letters before the buffer pointer.

8..

Any conditional can be reversed by preceding the " by a
colon modifier.
Thus O:"E fails.
·

#is an exclusive or arithmetic.operator.

(·turns off the colon and @-sign flagsw
The matching )
turns them on if and only if they were un before the (,
but will never turn them off.

11.

: and @ are both general command modifiers.
Preceding a
command by either or both of these modifiers changes the
command into a different (but usually similar) command.

clear~

as .,.+n if n is positive.

screen when executed.
the~evice

Page 59

MIT TECO (cont.)
12. n= outputs the va1ue of n in the current output radix. The
output radix is initially decimal and .is kept in a special
Q-register whose name is .. E (MIT TECO has lots of special
·Q-registers of this form).
13. @> is the same as > except that the iteration count is not
decremented.
!<!@> is a good way to jump back to the start
of the iteration without decrementing the count. The reason
for the angle bracket in the comment is so that if the
iteration should end earlier (say by a ~; command), then
TECO won't branch to this > , since it will see the <
(MIT TECO doesn't ignore angle brackets found in strings
or tags).
14. n:A appends n lines of the file (but won't append beyond a
paqe boundary) .
15 .. :C is like C, but returns -1 ordinarily.
It returns ~ if
C without the colon would cause an error.
:J and :R act.
in a similar manner.
(:D doesn't seem to be implemented
though.)
16. EDfile$ deletes the specified file.
17. E[ pushes the input channel if any.
Saves the current input
file and position in it, or saves the fact that no file is
ope.n.
18. E] ·pops the input channel.
If any input file was open,
the rest of it is flushed.
19. E\ pushes the output channel.
2~.

Et pops the output channel.

If an output file is open, it
is closed without being renamed.

21. m,nFt@ returns two values, which are m and n in numerical
order.
22. FtY returns a value saying how many args were given it
(in the low order 2 bits). The 4's bit is turned on if a
: preceded the FtY and the B's bit is turned on if an @
preceded the command.
23. Within a macro, FtX returns as its value, the arguments
that were given to the macro . . This command is frequently
followed by a FtY to determine the number and nature of
those argumentso
24. FQq returns the number ·of characters in Q-register q.
-1 is returned if the Q-register contains a number.
25. m,nGq inserts only a part of the text in the specified
Q-register. Specifically, the range from m to n-1
inclusive is inserted.
In this article, we have only been able to touch upon a very
brief sampling of the commands in MIT TECO. We will continue
this article in the next issue of the Moby Munger.

Page 60

MULTICS TECO
~ULTICS TECO is a.version of TECO that runs under the MULTICS
~perating system on a Honeywell computer~
It ·is of interest
to us because it is based on an early version of MIT TECO and
consequently, its advanced features highlight those features
of MIT TECO that were deemed most useful. Multics TECO was
first written late in 197~ by Richard H. Gumpertz, then working
for MIT. Later, Peter Bishop, also working for MIT became the
maintainer..
In 1972, David Levin at Honeywell became the
principal maintainer and augmenter unt~l 1977 (or 1978) .
In
1976, Richard G. Bratt· added a slew of features derived from
ITS TECO.
In additiori, in the mid 1970's, several other
people at the Cambridge Information Systems Lab (the Multics
System Software arm of Honeywell) , such as Mike Grady and
Larry Johnson, added certain additional features.
TECO's status in Multics is that of a 'tool'; it is not
part of. the standard service system and is used primarily by
sophisticated systems progranuners.
My thanks go to Bernie Greenberg who provided me with
the information for this article which comes from Publication
AZ,03, "Multics System Programming Tools".
(Incidentally.,
Bernie is the main implementor _for a Multics version of EMACS
written in LISP.)
A significant difference between Multics TECO and the
TECOs that you may be familiar with is that Multics TECO does
not use ani special characters like ESCAPE or control characters.
This is probably because early Multics users ran Multics on
terminals.that could not transmit these Gharacters~ Consequently,
all strings are essentially @-sign modified, although no @-sign
is required. Thus, to insert.some text, you type I/text/ and to
search for a string, one would type S/string/ where "/" denotes
any non-alphanumeric character.
In addition to the /text/
form of a quoted string, the construct 6q may also be used, where
q represents the name of a Q-register. Thus SQA searches for
the string that is stored in Q-register A. Conunand string
execution is begun by ending a line with a dollar sign ($)
(followed by a newline). Other than these differences, Multics
TECO ·looks very mubh like traditional TECO.
The following features are extensions to 'traditional' TECO:

There are 95 Q-registers, one for each printing ASCII
character.

I is used as the boolean 'or' operator.

Input and output is a bit different because complete files
are read into the text buffer so there is no need for
p·aging; consequently no need for the A, P, Y, or PW
commands. A file is read in with the EI/file/ conunand
and written out with the EO/file/ conunand. EO can also
take a numeric argument or argument pair representing a
range of characters to output (as in the T command) .

m,nS/string/ is iike nS/string/ ~owever only m lines
are searched.
(This is a bounded search command).
If m is negative, previous lines are searched.
If n is
negative, the search proceeds backwards and a successful
search leaves the pointer at the beginning of the string
found.

rage 61

N/string/ searches for· a regular expression, ·string.

m,n= types out the two numeric· values separated by a space.

n:= types out the value of n in octal.

8.
9.

EQ (External Quit) exits TECO.
:Iq/string/ or n:Iq inserts th~ specified text into
Q-register q. ·

l~.

n:L is equivalent to nLR (i.e. it moves to the end of
the line rather than the beginning) . Note that in
Multics, lines are ended by a single newline ~haracter.
n:J, n:C, and n:R are like nJ, ·nc, and nR except that
.if the text buffer pointer would ·be moved out of bounds,
no error is generated and the text buffer point is set
to B or Z as appropriate.

11.

12.
13.

m,nUq is equivalent to nUqm .
m,nPq is like m,nXq except that the text is appended to
the specified Q-register·.
(The letter P was free in this
version of TECO, however, for TECOs that do.P.aging, I recommend
the :X command to be the append to Q-register command.)

14.

:< and n:< are like < and n< except that errors that
occur within the iteration group just terminate the group
and the > returns a value.
-1 is returned if no errors
·occurred, ~ is returned if the iteration terminated due to
an error.

15.

:; is like ; except that the sense of the test is reversed.
n; exits the current iteration if n<~.

16.

If the target of an 0. command is not found in the current
macro level, the previous macro level is searched. This
continues back to command level. Although a GOTO can exit
a macro, it cannot be used .to exit an iteration.
(A funny mixture of extension and restriction.)

17.

:Mq, if issued from within a macro, causes a return from
that macro to occur when macro q is returned from.

18.

EM/file/ executes the TECO commands stored in the specified
file.
String arguments may be passed to a macro. The general
call to a macro has the following form:

19.

m,nMq/arg //arg //arg / ...
1
3
2
The command :Xq, if issued from a macro, puts the next
argument ~nto the specified Q-register.
(The first :Xq
gets the first argument, the next :Xq accesses the second
·argument, etc.)
The macro and the caller must agree on
the number of arguments, since when the macro is exited,
control resumes with the first character ·after the last
string argument that had been fetched.

Page 62

2JL

m,n"E .... ' executes the commands if m=n ..
m,n"N ... ' executes the commands if mfn.

21.

"M/string/ ... ' executes the commands if the specified
string appears in the text buffer just to the right of dot.
:"Mis like "M except the sense of the test is reversed.

22.

:' transfers to the next'. This is tiseful in creating
if-then-else statements, e.g. n"Xthen:'else' .
VW reads a character from the terminal and returns its
ASCII value.
(We are familiar r.-vith this as a CTRL/T command.)

23~

24.

:VWq reads a line of text from the
it ~nto the specified Q-register.

25.

EC/string/ passes the string to the Multics request
processor (conunand procepsor) for execution.

26.

EAq/string/ passes the specified string to the command
processor's active function application entry. The result
is returned ·in Q-register q.

27.

nA returns the ASCII code for the (.+n)th character in the
buffer.
lA gets the value of the current character (the
one just to the right of the text buffer pointer) .

termin~l

and puts

(This is compatible with MIT TECO as opposed·to TECO-llo)

. 28.

: \ i s like\ except that radix·B (octal) is used.

29.

m,n\ inserts the decimal interpretation of m into the text
buffer (before dot) . It is padded ·on the left to be at
least n characters wide.

3$L

$ (embedded dollar sign) throws away any numeric arguments.

31.

Special loop facilities are provided for throwing and
catching values. F< and > define an iterative group like
< and ~ .
From the time that the F< iteration group is
entered until the time it is.exited, F< sets up a handler
to "catch" values thrown by the F;/label/ request.
If no
F;/label/ request with matching string argument is executed.
before the F<!label!
group is. exited, the iteration
group returns -1 as a value.
If, however, an F;/label/
request is executed (where the /label/ match~s~drie that
·immediately follows the F<, i.e. F<!label!), then the
execution of all macros and iteration groups encountered
since the F<!label! is abandoned, and the F< iteration
group returns the numeric argument of the F; request as a
value.· :F<!label! is like F<!lapel! except that if an
error is encountered·during the execution of the :F< group,
the latter returns ~ as value~ These requests provide a
method of exiting several nested loops at once. Execution
of a F; command terminates the F< loop as well as any
contained loops·.

32.
33.

The Q-register '' always contains the last quoted string
that has been referenced by TECO.
:T/text/ types out the specified text on the terminal.

34.

A period at the end of a digit string is an "octal point".

TECO NOW RUNS ON YOUR HOME COMPUTER:
. TEC65
TECO ON 6502 MICROCOMPUTER SYSTEMS
A subset ot' TECO developed to run on
s ys t e Ms LI s i l) ·a t he 6 5 0 2 M i c r 0 -

Mi c r 0 c 0 M IHI t e r

processor· chi p i s ci v ci i 1c-1b1 e for ~.; y ~'>'l:. e Ms ·LI s i n •3
T I M,
S '( M, c-1 n d ~{ I M M o n i t o r s •
I t i s <:1 l s o s o rJ n
to be available for APPLE II systeMs.

TEC65 <as i t is na~ed> w~s written by
of
It e n v e r , Co 1 o r <:1 do wh o b e c c' "' e
faMiliar with TECO iMpleMented on a PDP-10.
TEC65 requires Just_ under 4~{ bytes of RAM ..
T o t c1 1
RA M r e q u i r e d d e p e n d s o n t he <:1 Mo LI n t o f
text you wish to handle at one tiMe.
SYM ~nd
KIM versions allow saving and loading of text
via the cassette pbrts provided.
The new
APPLE II version will Mctke use. of APPLE II

L. c1 r r y F i s h

Iii sk hc:1rdwc-1re.

A catalog describing this and other
6502 software is availC1ble for $1.00 froM the
6502 PROGRAM EXCH4NGE, 2920 W MOANA, RENO, NV
89509.
Ilavid f'. Mc:1rsh
Vice President
The 6502 PrograM Exchange
2-920 Moana
Reno, NV 89509
Remarks from the editor:

Yes, it's really.true! To quote from the catalog, "There is
no reason why you shouldn't have one of the finest texteditors on the market today".
I've got a copy of the TEC65 manual, and indeed, they have
a real TECO, although i~s a subset of the traditional TECO
for the -1.0.
The following commands are implemented:
C, D, FSa$b$, Gq, Itext$, J, K, L, Mq, T, Xq, =, < ••• >, Stext$,
;, Ntext$, *q, EX, ., H, B, Z, and TABtext$.
There are also a few other I/O commands that are riot compatible
with .traditional TECO, so I won't go into them here.
Q-register names are limited.to .0 through 9. There are also
the usual immediate action commands such as rubout and CTRL/U,
as well as LF which performs.an effective LT if typ~d as the
first character after TECO's prompt. Also CTRL/G which
executes an effective .0LT and ESC which executes an effective
-LT. The match control construct CTRL/X is also permitted in
search strings.

Page 64

QUESTIONS AND ANSWERS
This c·olumn will attempt to answer questions_ .submitted by
readers on any TECO-related topics.
Querie~ should be submitted
to the editor, Stan Rabinowitz, 6 Country·c1ub Lane,
Merrimack, NH 03054.
Q4. by Ronald J. Price (U.S. Environmental Protection Agency,
Houston, Texas)
I am presently using version 3 of DEC's OS/8 TECO, but I
have version 4.
How can I get information concerning the
differences in the two versions?
A4. y4 came out with a new release of OS/8 and has no new features
but contains bug fixes.
You should be using V4 rather than V3
because it is solider.
In fact, you should get the latest
version (VS) since it has considerably more functionality than
V4 ..
. Q5. by Ronald J. Price
I am using OS/8 TECO on a PDP-8/E with 32K memory.
How can
I increase the Q-register size from a maximum of 2000
~haracters to something larger?
AS. On BK machines~ you get 2000

characters in Q-register space.
10
On machines with 12K or more, you get a maximum of 2944
10
characters in Q-register space (although no single Q-register
may contain mor~ than 2047 characters) .
I know of no way
to·increase this limit of 2944.
Allowing Q-register space
to cross field boundaries is a ~ery tricky operation;
·
·although with 32K you have lots of room to write tricky code.
If any'one can provide a patch to OS/8 TECO to allow
expanded Q-register space, please write in.
.. ...,.. ..
, ..
'Q6. by Dr. R. N. Caffin (Division of Textile Physics, C.S.I.R.O.,
Ryde, N. S. W. )
_

I have several questions on VTS2 support under RT-11.
The Backspace (BS) key is meant to move the cursor to the
end of the current line.
However, under certain conditions
after editing within the line, it fails to do so: it ends
up some characters short of the end. Does anyone have a cure
for this?
A6. I have never encountered any problems with the BS function.
Please check that you have an uncorrupted version of VTS2
(better yet - switch to VTEDIT).
If the problem recurs,
please try to send circumsta~ces which can be duplicated
which cause the error to happen.
Note that the implementation
of this function is not too swift; it simply executes the
TECO commands L2R.
This means that if the line ended with
something other than CARRIAGE-RETURN / LINE-FEED, typing the
BS key may not do what you expect.
This could happen if the
line ended with FF ·or end-of-buffer.

Page 65 ·

Q7. by Dr. R. N. Caffin
Could we arrange to have future releases of VT52.TEC to
omit the screenful of keypad layout, assume that all
necessary files (except for EDIT.TEC and the text file)
are on SY:, and to allow lower case handling automatically?
A7. The source of VT52.TEC is distributed in all releases.
This allows users to make their own custom modifications.
We can't please everyone, so we release it with the
options that we feel are most useful.
You should feel
free to make your own modifications (but be sure to document them; and don't send in any bug reports unless the
bugs occur in the macro a~ distributed by us) .
QB. You mentioned in MOBY MUNGER #2 the existence of VTEDIT.
An interesting extension of VT52, with its word orientation,
but how does one get a copy?
·
AB. By the time you read this, VTEDIT should· have been submitted
to the DECUS li.brary, or will be shortly. Wait for announcement of availability from DECUS, then order it from DECUS.
Copies were also given out at the last DECUS symposium.
Q9. What is the chance of including officially, in VT52, the
command CTRL/P for "scroll one page-full forward" and (say)
CTRL/Q for "scroll one page-full backwards"? It's a very
useful facility for browsing through a long listing or a
piece of text.
A9. The latest version of VTEDIT (Vl.00) already has this feature
(but with a different syntax). See elsewhere in this issue
for the functionality of the latest version of VTEDIT.
QlO. Is anyone working on an extension to VTEDIT which will allow
direct preparation of text, bypassing the use of RUNOFF?
The ·main features required are the automatic adjµstment of
words in a paragraph following an insertion or deletion,
right margin justification and listing with page numbers.
AlO. I know of at least· two people at DEC who have macros that
do this kind of thing.
I can only hope that they will
~ubmit them to the DECUS program library.
Qll. Is it possible for TECO to perform direct output to another
terminal or line printer?
All. Yes. Unless the operating systems denies you access (say
if the line printer is assigned to another job) there is
no reason why you can't perform an EW command ·to such a device.
If you are using TEC0-11 and you are in the middle of an EB,
you can open the line printer (or terminal) for output on
the secondary· output channel by issuing a command such as
EA EWLP:$ and then issuing P commands.
The EW$ command then
switches you back to the primary output channel.

Ql2. by Dr. R. N. Caffin

Page 66

In humble ignorance_, what is the meaning and origin of MUNG?
Al2. MUNG is a recursive acronym for Mung Until No Good.
This is
what you typically do to files with TECO.
I believe the
word originated at MIT but am not sure. Better information
concerning the origin of this word would be appreciated.
Ql3. by Chester Wilson {Charleville Australia)
I have the DECUS TECO release for RT-11 and am having a
difficult time working out such things as chan~ing the
default ET and ED values.
Is there a simple way to do this?
Al3. RT-11 TECO V34 (soon to be available from DECUS) supports
user initialization via a file called TECO.IN!.
The TECO
commands in this file will be executed before TECO starts up.
Ql4. by J .• M. "Pete" Asensio
AS newcomers to TE~O, we find it somewhat difficult to
decipher involved programs, such as the scope macros, when
no comments are present in the distributed versiori.
Is it
possible to obtain a commented listing of VT52.TEC or
VTEDIT.TEC? We wish to patch these macros for use with
no~-DEC terminals, specifically the Infoton I-1~~.
Al4. no answer supplied.
The authors were "too engrossed" in
getting this marvel to work that they didn't bother to
include any comments. Anyone who ·has conunented it should
send us a copy. We hope that other readers who send us
useful TECO_ programs will properly comment them.
Ql5. by Nick Carr (DEC Europe)
My main use. of TECO is with RSTS/E. Unfortunately, these
scope macros run rather ~lowly with RSTS/E.
It is possible
to type ahead several characters (apparently between run
bursts) and -then TECO will suddenly catch up.
Would it be
feasible to increase the speed of these macros, possibly at
the expense of some functionality?
Al5. We are always trying to increase· the speed of VTEDIT.
Suggestions for speed improvement are always welcome.
If
the latest version of VTEDIT seems too slow for you, you
might try switching back· to VT52.TEC - it should run faster
but has considerably less functionality.
It is always
possible for you to go through VTEDIT removing features that
you don't want.
You can also shuffle the code around.
Putting those features that you use most frequently at the
beginning wou1d speed these features up.

Page 67
Ql6. by H. s. Hopkins Jr (General Latex and Chemical Corp.
of Ohio)
Let's say I'm in the .middle of the third page of a 5 page
TECO source, and want to bring in (insert) several lines
or all of another source file.
I keep lostng 1) the rest of
the 5 pages, 2) my place in the s'ource, or 3) everything
except the backup. file.
Is there a convenient way to make
this type of insertion?
Al6. If you know in advance that you want to make these insertions,
you should open the file containing the material·to be inserted first.
Read the file in, ·store the material to be
later inserted into a Q-register, and then begin your edit.
If you are using TEC0-11, another solution is possible (this
method is not available with OS/8 TECO or TOPS-1~ TECO) :
Store away the current text buff er (HXAHK) , open up the other
file using a secondary input stream (EPERfile$Y), locate
the material to be inserted and store it away in some Qregister, delete the unwant~d text and switch back to the
primary input stream (HKER$GA) and restore the text buffer.
Then resume your editing.
Ql7. Conversely, I would like to take 15 or 20 lines from page 4
and put them somewhere on page 2. How can this be done?
Al7. Some TECO variants allow reading the entire file into 'memory
to let you do editing like this.
If you don't have such an
editor, there are several ways you can proceed. The simplest
is to move the text into a Q-register, finish your edit
without leaving TECO, and then do another EB and insert the
text where desired. Another solution (which only works if
·you have enough memory and the two pages are near each other)
is to read several pages into core at once; enough to include the two pages in question. When doing this, don't
forget to re-insert the form feeds that the Append command
removes when appending pages together. For example, if
you were on page 2, instead of doing a P when you were done,
you would issue a AZJ12I$A command. Pages 2, 3, and 4 would
now be in the text buffer and you would have no t~ouble
moving text from page 4 to pag~ 2. Another solution, for
longer shuffles, as you say later in your letter, is to break
the source up into severa1··files, edit these as necessary,
and then recombine the files in the desired order.
Ql8:. by Ray Newman (North Brisbane, Australia)
We have· changed the effect of the CTRL/Z command (in the
VT52 macro under RSTS/E) to do an EX.
Unfortunately, this
will, if and only if the private default run-ticie system is
BASIC,· leave the terminal with the echo turned off. How can
we get around this problem?
Al8. This problem has been solved for you in VTEDIT. With that
macro, you use· the ESC CTRL/Z to perform a clean exit. You
probably could have solved the problem by explicitly turning
echo back on via the command (-8-l)&ETETEX .

Page· 68

by Ian Calhaem {Board member, bECUS Australia)
I have made modifications to VT52.TEC (as supplied in
RT-11 V3) to enable use with a VTlf:l_Y1.
How can I ensure
that a RESET escape sequence is output after the command
EX$$ ? How can I convert from 80 to 132 columns and
reverse?

19. Get and use VTEDIT when it becomes available. This is an
expanded version of VT52 which fully supports the VT1.Y1.Y1.
This macro supports 132 column mode and other VTl,0~ ·
features (for example, it displays your 'cut buffer' in
reverse video.
To quit editing, inst~ad of going back to
TECO and issuing an EX command, you type the ESC CTRL/Z
command to VTEDIT. VTEDIT then sends the proper escape
sequences to your terminal (if it ~s a VTl~f:l) to put it
.back in a nice mode, and then exits TECO.
)20. Can RT-11 TECO V28 be made to run as a foreground job?
I would like to run in foreground and background at the
same time.
A20. Yes.
It is very easy to get TECO to run in the foreground
under RT-11. All you have to do is link together the
object modules (TECO.OBJ and TECOIO.OBJ) using the
/FOREGROUND switch.
(You should also specify a l~,0-word
stack.)
The result is TECO.REL which can be run in the
foreground via the . FRUN TECO command·~ Mark Bramhall
reports that there were some bugs in version 28 of TECO
for RT-11 which might cause problems with running TECO in
the foreground.
All these bugs have been fixed in V34.
In fact, running TECO in th~ foreground under V34 is even
simpler, since no re-linking is necessary1 a prelinked
TECO.REL is included in the V34 TECO kit.
Q21

by Ray Newman (North Brisb.ane, Australia)
Under RS~S/E, two of the internal flags behave in different
ways depending on whether the user is privileged or not. On
scope terminals the ET flag and the EU flag are both f:1 for
non-privileged users.
For privileged users these flags
take their correct values.
How do we get around this?

A21

You are using an old version of TECO.
This problem has
been fixed long ago.
The TECO that was distributed with
the 6C release of RSTS/E does not exhibit this failure.
Use that version, or use V34 when it becomes available -from
the DECUS library.

Q22. by Stanley Rabinowitz, DEC Merrimack·
What is the origin and derivation .of the term "Q-register"?
A22. No answer supplied.

Responses are solicited.

:Page 69' ,···'

THE SEWERS OF PIING POONG
Ned Freed has written in to.tell us about the many TECO programs
that he has. These were written for TEXAS TECO version
124(344) .
I hope that he will submit them to the DECUS program
library.
In the mean time,. he will send them to anyone who sends
him a magtape. The two he sent to me were too big. to print.
One, called TECINI, was 4 solid pages.of TECO code and is a
utility to allow quick construction of good TECO.INI files.
It allows TECO novices to have very sophisticated macros at their
disposal, and for experts to genera~e them more quickly.
The other program is called SEWER. It was originally written
by Darryl Morrell and modif~ed by Charles Carvalho.
I was quite
impressed by this game.
It is a mini-ADVENTURE and was almost
5 pages long, inclu.ding such informative messages as YOU ARE
SITTING IN YOUR LITTLE MUD HUT IN PIING POONG EATING YOUR BOWL
.OF RICE WHEN ALL OF A SUDDEN A TYPHOON BLOWS IN. THE RESULTING
FLOOD WASHES YOU OUT OF YOUR HUT AND INTO ... THE SEWERS OF
PIING POONG, and YOU ARE IN AN ANCIENT CATACOOMB WITH PILES OF
BONES EVERYWHERE. A LARGE VAULTED PASSAGE WEST IS NOW BLOCKED
BY A CAVE-IN. THERE IS A SMALL HOLE IN THE EAST WALL. THE
SKELETONS ARE WEARING VARIOUS PIECES OF JEWELRY, ... ,and
YOU ARE ON THE SOUTH EDGE OF A STREAM OF SLUDGE ...•
The object is to get out of the sewers.
Ned's ·address is:
Ned Freed
Seaver Computer Center
The Claremont Colleges
Claremont, CA 91711

HELP PROGRAM FOR RSTS/E
Ray Newman (of North Brisbane, Australia) sent along· the
program HELP.TEC (printed on the next p~ge). This program
responds to the RSTS/E CCL command HELP.
If there is no
following argument, it lists topics ~or ~hich help is available .
.Arguments that follow the command are used· to select information
from one of two help files and display it·on a VT52.
It first
searches the users PPN for the file HELP~HLP. ·If it· fails to
find it there, it searches [1,2]. The format of the file is:
One topic per page and all topics and sub-topics enclosed in
matched braces . . Ray s~ys he is indebted to RTll for the idea,
but not the code which is of the home-grown variety.

HELP.TEC
-· ·- _, -· ...............
.......-..
-···---···~

'"'"""" ,. .. ,.. •••• .. """ ..' •••-u••·~••"••••-•-··--·-.-•••••-""""'-w••·•-•-"•"••·•-"•·-•••

ResPonder to the HELP CCL

Hoo

. . . Moo

o•o

PO•

.... ""'

,._

Set

U"o

-~ · - . . . . . . . . 0 . . .

• . . - - o00· . . . .

'o 0

Woo - ·

the various Q

ETt1ET 27

@~US\UT

0•

QT~T

H""

. . . HOO-·

0Po

... - ·

- · W"

t
!
'

re~s
ET&C1~->~T\

@'.... UC\{}\

ZU1 OA Q1J :@SI I
•r Q1~.-1XD .,zxA Q1,ZK ,
OU8 OU9 -1ELJ -l~X @~UE/

@MU~\ZJ

@EW// G*

Close off command file
and cht.-'ck for corr·ect CCL
@EI//
J ::@S/HELP/ 1 F·@~A/?lllesal CCL command
I ..... C'
!
Clean u2 and stur·e comm<:~nd
0 ir + K "~ D Z ..J If'.( / / J
! r~EM ! : ~ (~s;~cs/ 11 T -D @o ! r~EM ! ;

< @S/

/;~DJ

HXA HXD HK

Tr~

for Private helP file
if that f~ils ~o straisht
to the s~stem file c1,2J

: C~ErUHEL.F' + HLP / • F
! ~rl' SF I I.. ! [ [I J r1 1 U9
1 f~ Et"~ I [ 1 , 2 ::J HE L F' • HL F' I

• F @..... f-l /

'f'N<J he-lP file available
I

"'l·~

I I
y
Check ~or command if none
then Just list topics

OB •N @O!TC' .. M7. •r. @,...A/
Help is avsil~ble for the
I
0 lJ 1 ! TC ! < J D @S / ,., S /
1 ou1 / < ·-· ( o1 & 1 ~j > ;

a9 ·c 1us &01svsrIL!
10?
!
OU1

D , • -· 1 T • -· t %1 ·· 6 0

(~,...A I

' @-Ai
. '""C'
The main loop starts here
< OJ < !ARRGH1 -1'""X @S/{/;

C~~~·<J. , ZK

B, (.~2K

'TYPE 1 T

@'"' r.·J

< @S/ ..... E:GC/ t

I ET.&~; 12
HT fr'E-·V
/ ! E T &5 1. 2 ! 1. • N 9 2 Ms

•G

toPics:

er· r. . 1 ./ :1. ~' :J. > . .~ N,.,, _.. ; Y >

+ U2
1 U4 < l@S/""TGC/ °F @,...A/
"f Co r· ruPt HE LP+ HLf' file
I ..... c I
1 %4 0 -· 1 A .... 1 2 5 I E ()-·· 2 %4 • E + lJ 3
!.C 0 NT !
Q ~~ J 0 '"· X : : @S rT QB I
" T (~SI'"' S /
MZ 11 N C.B··-1vZI< B,Q2K J (10!ARRGH' "

: GE

followi~~

~N~_;

*> Q9 •[

.... c

@O!SYSFIL~

XNo help for /

!GD

·"·[I

@~A/

:::. !

* ET&512 *! 1 N /'2MS 74MS 91MS ,.
1

(.~ 2 .J
I

@O ~ 0 NT ! , I >
Q 2 , • -· l ! XE t @. '" U EI I

@,...Al

•E

t~ 3 ,J

Page 71

SQU.TEC • the TECO micro IQUiDher
SQU,TEC

•~he

TECO macro

SQ~i1~er

Mark Brarnjo,al 1

Thfa ~•cro w111 :take a Mice, readable TEC0•11 macro and sQuis~ it
10 that 1t is e1 1M•11 and•• feat•• poaa1b1e (&nd 11 completely
unreadable), SQU 11 distributed with the TEC0•11 kits and will
run under a~y operating system that aupoorts TEC0•11,
You invoke

SQ~:with

the MUNG commend,

For

e~emplel

MUNG SQU
SQU t~en enters 1nto a d1elo~ue wit~ you ea described below. All
auestion1
contain tne allo~ed 1ege1 responses enclosed i"
parentheses and the default you Qet if you Juat ty~e RETURN in
angle brackets, · In ti-I• diseu11ion below tne -term "1e)(iea1"
refers to e~araeter• that TECO would have 'interpreted •• TECO
com~sndsJ
111 e~araetera in1ide atri~g srgumeMts are not 1ex1ea1
characters, Fo~ example, the ABC of SABC$ ere not le~ieel
c na ,. a e t e r 1 wh ; i e t he, ·S a rtd; S a ,. e ,
S 1m1 1 4a r 1 y , t ne ABC a r1 d 0 EF o f
Ct FSI A8 CID EFI 1 re· not 1e >c 1c e 1 w1-i·; 1e the •, F, S, 1- 1, • nd $ are.
Delete CR/LF CV/N) <N>?

A Y response w111 d•lete •11 lexical RETURN& 1nd ~INE FEEDs.
tht1
usually ~~•ults 4n one very long "line", an answer of

As
N

is.

1ugge1ted Cbut, aee the Question below),

·Set 1 ine lengths .CY for 70,

Of'

lengtl'\)

Cf\J>?

Thia Question 11 asked if you reaponded N to tne ~revious
que1t1on,
If 1 coeitive reapo!"lse 11 oive,,, SQU will iMiti•11y
delete a11 lexical CR/LF•~ then, es a oost•oroees1, insert the
CR/LF ecmb1netion wherever it'• ~eeded so t~et ~o line is longer
'than th~ lenqth specified, Thia n~atly formats the sQuished TECO
ma e ,. o, · The ·mo 1 t
comm o,, 1 y u • e d 1 e Mg t "
reau1ts 1n • 1eni-th of 70,

i s 70

8 ·re• pons e o 1 Y
·

Delete blank·· 11~es CY/N) cN>?
Thi• Questiori is asked o"lY 1f .you respG,,ded N to both of the
Que a.t 1o,., a above.,
A Y an 1 we r re 1 u 1 t 1 i "' e 1 1 ·1 e )( 1cal b 1 a,, k 1 i ,., e 1'.
behig de1eted.
Delete le•ic11 TABS arid FORM FEED• (YIN) <N>?
Even

t~ouoh 1exiec1 TABa eMd FOR~ FEED• have mean1no to TECO CTAB
a ri. i n s e ,.. t , F 0 RM F EED o u t r> u t a a F 0 RM F EED .o n. t h e t e r mi n a 1 ) ,
acme TECO ma~ro writer• use them a1mo1y to fo~~et their m~ero
code,
Of eour1e, their ma~roa wori-t correctly execut• without
~emovino tho1e 1ewieal TAB• end FF1.
A Y re1pon1e w111 delete
all
lexical TABa·.,,d FORM FEED• •. If your meero f"UMI ~MIQUfshed,

I t a

rt i

you certainly wal"'\t to Ol"'\lwer N,
Delete commel"'\tl CY tor SP/TAB, N, or

1et)

cN>?

3QU.TEC • the TECO macro 1Quf 1her

PAGE

The t~ouble here f s t~et both comment• aMd 1ebe1s Cteos) use the
••m• conatruet,
To be able to differenti1t~ between the two,
tome tort of coding convention must be ueed. SQU wi11 cheek the·
Character immediately follOWiMQ the "1" th•t &tarta the COmmeMt
or label, If that ch•raeter metenes the o~e of t~e 1et of
"comment cherecterari s~ecified, SQU will delet~ the conat~uct,
U1•nQ e SPaee or TAB 1mmed1cte1y ofte~ the leading exclamation
point seema to be the moat 11e•ib1e tor formatting comments so a
Y re1~on1e re1ult1 1n the "commo"t character" set beinq S?eee eMd

TAB.

Wat~h

Prooreas (Y/N) cN>?

If vo~r term1na1 is supported by the W Cseope watch) commands,
thia QUe1tion w111 be eaked, At t~e coat of SQU runnino ~or•
a 1ow 1 v·, you e •,, w• t c n ft • P r o g re a i •
Ua e f u 1· to r
debug g i n g,
·f"I o t
for ~~oduct~oM 1au11h1Mg,.,
Automatic mode CY for

x, N~ o~ s~t) <N>?

Thia 11 t~e mo1t confusing aue1tio~. The problem here 11 that
no rm• 1 1 y wo M ' t to u eh e tu rec t e' ,.., in a i de 1 t r trig e r gum en t
1tPi"0•1 but the ~u eomma~d 4• freQuently u1ed to load TECO mec~o
cod• tnto Q•regist•r• and, of course, ~ou wiah that eode to 1110
be IQUi shed,
OM
the othet- :l"iarid,. 111 · •u eonHr1aMd1
aren•t
nece11ari1y loading TECO ~aero code, There e~e two for~• o1 the
•u eommandl •uax~~! end ••UQX•XM~. The rule SQU uses 11 t~•t
•"Y •u commend.that 41 of the ••UQXxx~x fo~~ end w~o1e de1imiting
cheraeter f 1 a ~•mber of the "non•1Quisheb1e •u
eo~mand"
eh1t-acter , set
shouldn.,t be 1ub.;1oui1l'ledJ
eve!"Y otMe,. •u
eomme~d's st~ing ergumeMt ia assumed to be'TECO ma~ro code ind is
aub•1Qui1hed, · The co~v•"t4on we've edo~ted 41 that the GUOtinQ
character ~X" 11 to be uaed for textu&l Ci.e., "o"•eode) ~u
commeMd · erguments,
Therefore,
e
V
response
1et1 t~e
"MoM•eau4~habte •u command".character aet to~.
Vou mov enter
a ri y 1 e t o f c h a r a c t e ,. ( s )· y o u we n t • Fo r e x em p 1e , • r e P 1 v o f XI \
meaMs that env ~u string delimited by a percent sign, 1111~, or
beck1la1h ch1Peeter should not b• sub•sau41hed, An N response
p1acea you •~ manual mode, Whe"ever SQU encounters a •u eomme~d,
1t will' ·~euse, ring the term1ne1'1 bell, and await a reaoonse.
A
Y ••V• to oo ahead a"d 1ub·~aui1h t~e etri"g 8!"uomentJ
an N
meana to 1k1p it,
Obviously, you went to atert coding you~
mac~o• aeeord1Mg to some eonve~t1oM 10 that you don't ~eed m•nual

.SQ U

mode,

A11ow adJac•nt

ESCe~e•

CY/N) 4N>?

The o~oeeas of sQu1ah1ng • maero can result •~ two adJaceMt
ESCece char~etera ·where there w11n't befora, For e•ample, if t~e
eommtMd IFS/ABC// w•a u1ed1 SQU w111 turn it 1nto FSABCS!.
If
the macro will be read vie aM ER command, yenked iMtO t~e ~ext
buffer, copied to • Q•reo11ter, and executed this po1e1 no
~~ob1ems,
OM the other hand, 11 the maero i• to b• EI•d, the
adjacent ESC1pe1.will tr~og~r comma"d ewecution ~remeture1v.
AM
N respo"ae w111 direct SQU to never generate adJaeeMt ESCapea
.... ....

L-. -- A.

Ao..

......

5QU 1 TEC • the TECO macro

PAGE

~quisher

73 :

F.i 1 a <~r , TEC>?

sQu 1 1 n o w a s k 1n <' yo u f o r

t no f 1 1 e
s Pe e 1 of 1··c a t. 4" "'
o f v o u r TEco
macro and 1ndic1ting t"e it w111 default the a~tens1on to ,TEC,
After your reolv, SQU wi11 take off and stb~t sau1shiMg t~e
mae,.o, When it'H done, it will simply l"etur"' z.·:.. TECo's * pro,,,pt,
You ean now write out the sau1shed mecro howev&r ·yo~ want.

If YOU

1nvok~

MUNG

SQU ••

sew, f 11 riem.

f t w1 1 1 e n t e r 1 t u d 1 a 1 o ou e f.o r t h e a.au 1 s td n g op t 1 o n s ,
b ut
w1 l 1
start aQuiahing 1mmediate1Y atter aakf ~g about a~Jaeent ESCapes
then return to. the * promr;)t,
·you e 1 n i n v o k e SQ U .1 p e c 1 f y 1 ri Q e 1 1 of
• w1 t c ti e 1 ,
T e -·a ~id t e h e.1 a r e 1

/0
IL.
/B

/T
/C

/W'.

IA
/E·
~ach

a o u 1 s Ii tn g
·1

t· i, e

opt 1on s

Oe1ete CR/L.F

Set 1 f Me 1 engU1a

i·

Delete blank 11Mes
.-D·e 1 e t e 1 e llC 1e a 1 T ABI F F s .
O~lete epmme~ts ·
We;t ch pro g "'"· s s ·
· Aut om at i c mode
Allo~ adJaeent ESCa~es

switch shou1d:be exDressed is /XIY where"~" 11 the switch
1s eitlier Y or N, A nol"l•stenderd 11~e 1enQth can be
e~pressed as /LIMl"I where "P"ln"
is some non•zero d1g4t strino,
Non•standard choreeter sets ce~ be soeef fied for the /C end /A
1witehe1, but tne eharacter ·set 11 limited toe s1ng1e ehera~ter,
For ewa~p1e, /CB* sets· the comment eHarecter set to "•" a~d /A11eta the noM•sQu~ahable ~u commeMd character set to "~".
end·

"v"

You

caM also supply SQU with
spec 1'f1 ea·t ions, rot" exampl el

both

output

8P"ld·

f 11 e

MUNG SGU,outf11=inft1

SQU w111 enter 1ts dialogue .for the IQuishing opt10Ma1 then wf11
1mm~d1ete1y do the squis~ing, write out the ~.~ult, ~nd e•it,
If you have spec~fied e.verythiMg SQU needs Can fl"IJ:>Ut file, aM
output f 11 e,. aMd all of tMe o~t i or11 as sw1 tehes), it wi 1 l do the
whole 1Qu1shino operat1oM and e>e1t without fu·rther· if"ltervention,

A co~ve~tio~ used at DEC Cend recommended by the TECO SIG) is
that if a TECO moero is fnteMde~ to be SQUiehed, .the uMsauished
ve~sioM
be g4ven e ,TES extens1on (TECO Soure~) a~d t~• aQu1shed
v er 11 on be g f ,v en a , TEO e >et e l"I s 1 o ~.

ff 1 e · w n h · 1 ·• TE C

must always be a runnable TECO macro Cor program),

e )( t e f"l 11 o ri

Page 74

~TECO-RELATED

DECUS SUBMISSIONS.

The following items of interest to TECO users are available
from the DECUS program library;
N.umber

Desc-r iption .

.DECUS-8-873

RSTS Terminal Monitor on a PDP-8
Written by Andras Nagy. Thts ~ackage
lets you connect together OS/8 arid RSTS/E;
Its goal is to synthesize the powers and
versatilities of the RSTS and OS/8
·
operating systems. The sdnrce language
consists bf ~ 1 1U~~x:tµ.:;r-e of assembly code
'. and TECO macros·., ' '
. .

DECUS-1,0-116

QED

Written at the University· of Utah.
·This is an?tmplementa~ion of the ·wellknown line-oriented ed"fto'r known, as' QED.
It is written in TECO ctnd"runs under a
slightly modified version of a 197,0 copy
of DEC's TOPS-lS TECO.
r i~
: ~ ~·

DECUS-1,0-148\
DECUS-ll-.S3 J

,~·

PDP-11/1,0 Loade.r
_
Written by H. ·L. Farhsworth and R. B. fleisher.
Can be ordered unde:r'~ither number.
Consists of two T:E!co j;.>,rograms· that all-pw
full duplex conve~sation bet~~en TOPS-l,0
and a PDP-11 termiriaf. · Also 'lets a PDP-11
pro~ram to be-assembled on the -1,0 and.
loaded di;eb~~y intd:PDP~ll mamory.
•p.

DECUS-11-36,0

RSX TECO with Buf f~red ~ypeout and VT52 Window
Support, plus TECO DOCTOR Game
.
Written by G~ EverhJrt. This· consists of
two parts. The ~irs~ pa~t is a video
editor (for a V'r 52) writ ten in TEGO. . It
runs underia modified version of TEC0-11
which is e~ciosed. 'That ver'sion of T;ECO
is now obsolete sinc'e the 'latest TECO~ll
for RSX supports ~T52s properly. The second
part of this submission consists of.the
well-known, DOCTOR program (ELIZA) w~itten
in TECO.
~tis s~ightlY"le~~ powerful than
the original LISP'program~
It allows the
computer to respond to your statements as
would a psycho-analyst~ Arid·att~mpts to
carry out a conve~~ation with th~ user.
Should run under most TECOs. What does
that suggest to. you?
1

-OECUS-11-265

DOS-11 TECO
Written by Glenn E~ethart.
It is a copy
of version 15 of TEC0-11 which has been
interfaced to a new I/O module-to allow it
to run under the DOS-11 operating system.

DECUS

DIGITAL EQUIPMENT COMPUTEH USERS SOCIETY
ONE IRON WAY, MR2-3/E55
MARLBORO, MASSACHUSETTS 01752

MOVING OR REPLACING A DELEGATE?
Please notify us immediately to guarantee continuing
receipt of DECUS literature. Allow up to six weeks
for change to take effect.
( ) Change of Address
( ) Delegate Replacement
DECUS Membership N o . : - - - - - - - - Name:---------·----Company:
Address:---------,

State/Country:··-----·-------Zip/Postal Code: - - - - Mail to: DECUS - ATT: Membership
One Iron Way, MR2-3
Marlboro, Massachusetts 01752 USA

.·