[9fans] pascal, TeX

[9fans] pascal, TeX

[erik quanstrom]

i think knuth is the counterexample that proves the point.
knuth lives on his own planet. i believe it's called
"the art of computer programming." ;-)

also, due TeX's pascal heratage, it has an ungodly number
of compiled-in limits.  i haven't looked carefully for a
long time, but it seems that the maintainers have just made
the limits really big or changed the code in some way. in
the day (working on 780s) it was always fun to recompile
TeX to finish a document (because you'd blown some limit
you'd never heard of before).

not to start a flameware (troff esentially has the same
problem), but TeX is a great education in Why To Hate
Macro Processors. sendmail does better, but at least with
TeX you can get something useful done at the same time.

erik


Douglas A. Gwyn <DAGwyn@null.net> wrote:
> "Thomas Bushnell, BSG" wrote:
>> "Douglas A. Gwyn" <DAGwyn@null.net> writes:
>> > D De Villiers wrote:
>> > > No Pascal implementation ? Pascal compiler etc etc ?
>> > Why?  What use would it be?
>> It would let you run Pascal programs.

> And what use would *that* be?
> Seriously, are there any major apps written in Pascal?

	TEX was.

[9fans] pascal, TeX

[Howard Trickey]

Knuth didn't trust Pascal memory management for some reason, so he avoided
using records and "new".  He did most data structures in huge arrays of
integers with punny macros to fake up data structures inside them.  It was
Pascal with Fortran for data structuring...

> knuth lives on his own planet. i believe it's called
> "the art of computer programming." ;-)
> also, due TeX's pascal heratage, it has an ungodly number
> of compiled-in limits.

Re: [9fans] pascal, TeX

[Thomas Bushnell, BSG]

howard@research.bell-labs.com (Howard Trickey) writes:

> Knuth didn't trust Pascal memory management for some reason, so he avoided
> using records and "new".  He did most data structures in huge arrays of
> integers with punny macros to fake up data structures inside them.  It was
> Pascal with Fortran for data structuring...

Knuth's reason for this was not a lack of "trust".

Rather, he wanted to make TeX portable to languages which don't have
dynamic memory allocation; he wanted it to be reasonable people to
write simple automated translators from Pascal to languages like
FORTRAN, and so he avoided certain language features in Pascal that
are lacked in the other languages he was concerned about.

Thomas

Re: [9fans] pascal, TeX

[George Michaelson]

> not to start a flameware (troff esentially has the same
> problem), but TeX is a great education in Why To Hate
> Macro Processors. sendmail does better, but at least with
> TeX you can get something useful done at the same time.
>
> erik

I think a macro processor remains a fundamentally good thing. Else
how could Bourne have programmed his shell in Pascal?

The problem is divergeant Macro processors. better to have one and
make it work well. Isn't that what hit troff/eqn/tbl in the end?

I think the real problem for that family was the non-obvious cross
relationships of the *orderings* of calling the submembers. if you
want to nest equations in a table thats different to nesting tables
in a document with equations sometimes. Certainly trying to use eqn
to get giant curley brace marks into tabled telephone lists beat me.

(I was emulating a manually typeset phone list. you cannot recreate a
 craft like that with machines, Knuth not withstanding. I tried both
 TeX and Troff and troff won the day, but it was never quite as nice as the
 original.)

cheers
-George

Re: [9fans] pascal, TeX

[Alexander Viro]

On Fri, 14 Dec 2001, George Michaelson wrote:

>
> > not to start a flameware (troff esentially has the same
> > problem), but TeX is a great education in Why To Hate
> > Macro Processors. sendmail does better, but at least with
> > TeX you can get something useful done at the same time.
> >
> > erik
>
> I think a macro processor remains a fundamentally good thing. Else
> how could Bourne have programmed his shell in Pascal?

ITYM Algol-68.

Re: [9fans] pascal, TeX

[George Michaelson]

> > I think a macro processor remains a fundamentally good thing. Else
> > how could Bourne have programmed his shell in Pascal?
>
> ITYM Algol-68.
>

correctamundo. my call by reference to the name failed to get the right value.

God, I never came to terms with that language. anything beyond a 3 line
hello world was beyond me. Still, that two lines more than the APL I wrote.

cheers
	-George
--
George Michaelson       |  APNIC
Email: ggm@apnic.net    |  PO Box 2131 Milton QLD 4064
Phone: +61 7 3367 0490  |  Australia
  Fax: +61 7 3367 0482  |  http://www.apnic.net

Re: [9fans] pascal, TeX

[Boyd Roberts]

George Michaelson wrote:
> I think a macro processor remains a fundamentally good thing. Else
> how could Bourne have programmed his shell in Pascal?

Bourne programmed his shell in SHELLGOL; one of the true ghastly
abuses of the preprocessor:

    IF foo == bar
    THEN
	baz();
    ELSE
	boom();
    FI

And the only way to modify the Bourne shell was to stick with the
SHELLGOL, because at least it would maintain a consistant style.

I think laTeX is the definitive answer to why TeX is a poor
piece of software.

Write TeX or laTeX is like writing PostSript and I don't want
to write any of them.  troff, for all it's faults, I prefer.

Re: [9fans] pascal, TeX

[George Michaelson]

> And the only way to modify the Bourne shell was to stick with the
> SHELLGOL, because at least it would maintain a consistant style.

Well, thats the *good* way. there are many bad ways. I wonder what
freudian slip made me forget it was an ALGOLoid? Must be memories
of failing to chose that language course at Uni. Co-incidentally
York University had a truly wonderful sculpture collection, and one
piece outside goodricke college is aptly named 'algol' -It seemed to
fit given the Wirth focussed department of that time. (its a two-part
sculpture, algol being a binary star, but it fitted for -60 vs -68)

>
> I think laTeX is the definitive answer to why TeX is a poor
> piece of software.
>
> Write TeX or laTeX is like writing PostSript and I don't want
> to write any of them.  troff, for all it's faults, I prefer.
>

Me too, but as programme chair for AUUG I found I had no choice but
to hack the back-end postscript of the {troff,ms,pdf,tex} sometimes
to get eg page numbering to work. Might have been quicker to pay
a wage slave to re-copy all on a xerox with tip-ex over the numbers
and a golfball selectric to number them again.

-George
--
George Michaelson       |  APNIC
Email: ggm@apnic.net    |  PO Box 2131 Milton QLD 4064
Phone: +61 7 3367 0490  |  Australia
  Fax: +61 7 3367 0482  |  http://www.apnic.net

Re: [9fans] pascal, TeX

[Alexander Viro]

On Thu, 13 Dec 2001, Boyd Roberts wrote:

> Bourne programmed his shell in SHELLGOL; one of the true ghastly
> abuses of the preprocessor:
>
>     IF foo == bar
>     THEN
> 	baz();
>     ELSE
> 	boom();
>     FI

That's perfectly valid A-68, BTW.  And yes, degree of cpp(1) abuse is
horrible.  OTOH, I'm regulary seeing worse ;-/

> And the only way to modify the Bourne shell was to stick with the
> SHELLGOL, because at least it would maintain a consistant style.
>
> I think laTeX is the definitive answer to why TeX is a poor
> piece of software.

Now, that's hardly fair - you are blaming TeX for ugliness of SCRIBE
lookalike written atop of it.  AMSTeX is actually quite nice - for
mathematics-related text it wins hands down.  For anything CS-related
I'd pick troff, but...

Re: [9fans] pascal, TeX

[paurea]

Alexander Viro writes:
 > Now, that's hardly fair - you are blaming TeX for ugliness of SCRIBE
 > lookalike written atop of it.  AMSTeX is actually quite nice - for
 > mathematics-related text it wins hands down.  For anything CS-related
 > I'd pick troff, but...

I am wondering... you who criticize *TeX actually write real math
books (lots of formulas, commutative diagrams, integrals with strange
stuff...)  with troff and eqn?... I have found that there is where
they beat any of the other things I have tried (never tried troff though).

(please don't flame... )
--
                 Saludos,
                         Gorka

"Curiosity sKilled the cat"

Re: [9fans] pascal, TeX

[Steve Kilbane]

Gorka wrote:
> I am wondering... you who criticize *TeX actually write real math
> books (lots of formulas, commutative diagrams, integrals with strange
> stuff...)  with troff and eqn?... I have found that there is where
> they beat any of the other things I have tried (never tried troff though).

My experience with text processors is the same as with mail clients, editors,
browsers, etc. They all do some things better than all the others, and they
all do some things worse.

steve

Re: [9fans] pascal, TeX

[Douglas A. Gwyn]

paurea@dei.inf.uc3m.es wrote:
> I am wondering... you who criticize *TeX actually write real math
> books (lots of formulas, commutative diagrams, integrals with strange
> stuff...)  with troff and eqn?... I have found that there is where
> they beat any of the other things I have tried (never tried troff though).

I've used both.  Generally, TeX produces somewhat nicer-looking
output, but I find eqn|tbl|troff easier to use.  The DWB (troff)
suite is certainly more "toolkit oriented" and has a variety of
nice "small language" preprocessors for specialized tasks such
as drawing graphs, plotting, chemical symbols, etc.  Either
system is vastly better than something like Microsoft Word.

Re: [9fans] pascal, TeX

[Martin Harriss]

Boyd Roberts wrote:
>
> Bourne programmed his shell in SHELLGOL; one of the true ghastly
> abuses of the preprocessor:

I had heard that it was called BOURNEGOL :)

Martin

Re: [9fans] pascal, TeX

[Steve Kilbane]

Martin wrote:

> I had heard that it was called BOURNEGOL :)

It's called a lot of things, most of them not repeatable in polite
company.

Re: [9fans] pascal, TeX

[Douglas A. Gwyn]

Boyd Roberts wrote:
> And the only way to modify the Bourne shell was to stick with the
> SHELLGOL, because at least it would maintain a consistant style.

Finally David Korn converted it into ordinary C source.
(In case there is any confusion: I'm talking about the Bourne
shell shipped with UNIX System V, not the Korn shell.)

Re: [9fans] pascal, TeX

[Chet Ramey]

In article <3C18EF44.38244A15@strakt.com>,
Boyd Roberts <9fans@cse.psu.edu> wrote:
>George Michaelson wrote:
>> I think a macro processor remains a fundamentally good thing. Else
>> how could Bourne have programmed his shell in Pascal?
>
>Bourne programmed his shell in SHELLGOL; one of the true ghastly
>abuses of the preprocessor:

I still prefer `Bournegol', since he wrote adb in it, too.
--
``The lyf so short, the craft so long to lerne.'' - Chaucer
( ``Discere est Dolere'' -- chet)

Re: [9fans] pascal, TeX

[Dan Cross]

In article <28022.1008266218@apnic.net> you write:
>I think a macro processor remains a fundamentally good thing. Else
>how could Bourne have programmed his shell in Pascal?

I'm not sure that's a great example.  :-)  (btw- wasn't the Bourne
shell in Algol?)

	- Dan C.

Re: [9fans] pascal, TeX

[forsyth]

[-- Attachment #1: Type: text/plain, Size: 396 bytes --]

i suspect that's because the Pascal language took the novel approach of providing
neither guaranteed garbage collection nor a portable operator to free
memory without it.  there were several platform-specific methods, including
stack like heap allocation (which i think was used by one compiler).  `dispose'
ended up in the standard (but did not preclude implementing garbage collection).


[-- Attachment #2: Type: message/rfc822, Size: 2186 bytes --]

To: "erik quanstrom" <quanstro@speakeasy.net>, <9fans@cse.psu.edu>
Subject: [9fans] pascal, TeX
Date: Thu, 13 Dec 2001 12:51:37 -0500
Message-ID: <EPEOIODEGHAONLDKOGLHKEGPCBAA.howard@research.bell-labs.com>

Knuth didn't trust Pascal memory management for some reason, so he avoided
using records and "new".  He did most data structures in huge arrays of
integers with punny macros to fake up data structures inside them.  It was
Pascal with Fortran for data structuring...

> knuth lives on his own planet. i believe it's called
> "the art of computer programming." ;-)
> also, due TeX's pascal heratage, it has an ungodly number
> of compiled-in limits.

Re: [9fans] pascal, TeX

[Thomas Bushnell, BSG]

forsyth@caldo.demon.co.uk writes:

> i suspect that's because the Pascal language took the novel approach
> of providing neither guaranteed garbage collection nor a portable
> operator to free memory without it.

Umm, isn't "free" the portable operator to free memory?

Pascal implementations are free to use GC (and the language was
carefully designed to allow it, by prohibiting hidden aliasing), and I
think there have been some that use it.

Unfortunately, programs can't rely on GC, so they still have to call
free at the right times and do all the annoying and expensive
bookkeeping that GC is designed to make unnecessary.  You end up with
the worst of both worlds: no aliasing, and you still can't rely on GC.

Thomas

Re: [9fans] pascal, TeX

[D De Villiers]

According to my knowledge - Java was the first language that implemented
garbage collection...Neather C/C++ or Pascal (Delphi etc) have garbage
collection and memory must be manually located/delocated etc.

Regards,

Lennie De Villiers

--- Remove ~ and 9s from e-mail address to reply ---

> i suspect that's because the Pascal language took the novel approach of
providing
> neither guaranteed garbage collection nor a portable operator to free
> memory without it.  there were several platform-specific methods,
including
> stack like heap allocation (which i think was used by one compiler).
`dispose'
> ended up in the standard (but did not preclude implementing garbage
collection).

Re: [9fans] pascal, TeX

[Douglas A. Gwyn]

D De Villiers wrote:
> According to my knowledge - Java was the first language that implemented
> garbage collection...Neather C/C++ or Pascal (Delphi etc) have garbage
> collection and memory must be manually located/delocated etc.

Actually, garbage collection was implemented in some very
old languages such as LISP.

C++ does have what amounts to garbage collection plus other
programmer-controllable actions in its object destructors,
which are automatically invoked when objects go out of scope.

C has had garbage collectors implemented for it; for example,
somebody (I think it was Rob) made heavy use of gcalloc() in
the Blit family, both in the terminal's operating system and
in application programs.

The basic problem with garbage collection in C is nt in doing
it, but in doing it efficiently.  The main reason it is hard
is that C has excellent support for pointer arithmetic, so a
program can manufacture valid pointers in ways that are hard
to track, and a garbage collector can't (with a reasonable
amount of work) be sure when there is no pointer remaining
that could be used to access an object.  Nevertheless, there
have been some clever "good enough" garbage collectors
designed for use with C.  It's not obvious that they are
needed, since C has a long heritage of the programmer making
explicit calls to free() to deallocate dynamic objects.

Re: [9fans] pascal, TeX

[geoff]

Bourne's sh was written in a pseudo-Algol-68 by using the C
preprocessor, and indeed sh's input language is inspired by Algol 68.
If there hadn't already been an `od' command, the while loop would be
`while list do list od'.  The System V developers eventually took out
the Algol 68 macros in the source.  I think the first released version
of sh without the macros was the SVR2 shell.

Re: [9fans] pascal, TeX

[forsyth]

[-- Attachment #1: Type: text/plain, Size: 183 bytes --]

i've found that troff -mpm produces rather nice output,
even compared to TeX.
unfortunately it's not in the current distribution (partly
because the pm postprocessor is in C++)


[-- Attachment #2: Type: message/rfc822, Size: 2412 bytes --]

To: 9fans@cse.psu.edu
Subject: Re: [9fans] pascal, TeX
Date: Fri, 14 Dec 2001 17:04:29 GMT
Message-ID: <3C1A278D.6F117C50@null.net>

paurea@dei.inf.uc3m.es wrote:
> I am wondering... you who criticize *TeX actually write real math
> books (lots of formulas, commutative diagrams, integrals with strange
> stuff...)  with troff and eqn?... I have found that there is where
> they beat any of the other things I have tried (never tried troff though).

I've used both.  Generally, TeX produces somewhat nicer-looking
output, but I find eqn|tbl|troff easier to use.  The DWB (troff)
suite is certainly more "toolkit oriented" and has a variety of
nice "small language" preprocessors for specialized tasks such
as drawing graphs, plotting, chemical symbols, etc.  Either
system is vastly better than something like Microsoft Word.

Re: [9fans] pascal, TeX

[Ralph Corderoy]

Hi,

> > I've used both.  Generally, TeX produces somewhat nicer-looking
> > output, but I find eqn|tbl|troff easier to use.  The DWB (troff)
> > suite is certainly more "toolkit oriented" and has a variety of
> > nice "small language" preprocessors for specialized tasks such as
> > drawing graphs, plotting, chemical symbols, etc.  Either system is
> > vastly better than something like Microsoft Word.
>
> i've found that troff -mpm produces rather nice output, even compared
> to TeX.  unfortunately it's not in the current distribution (partly
> because the pm postprocessor is in C++)

What's tmac.pm and the pm postprocessor?


Ralph.

Re: [9fans] pascal, TeX

[bwc]

[-- Attachment #1: Type: text/plain, Size: 48 bytes --]

This is what's wrong with CS education today!!

[-- Attachment #2: Type: message/rfc822, Size: 2112 bytes --]

From: D De Villiers <~ddevilliers99@lando.co.za>
To: 9fans@cse.psu.edu
Subject: Re: [9fans] pascal, TeX
Date: Fri, 14 Dec 2001 17:37:15 GMT
Message-ID: <9vdbj1$ehk$2@ctb-nnrp1.saix.net>

According to my knowledge - Java was the first language that implemented
garbage collection...Neather C/C++ or Pascal (Delphi etc) have garbage
collection and memory must be manually located/delocated etc.

Regards,

Lennie De Villiers

--- Remove ~ and 9s from e-mail address to reply ---

> i suspect that's because the Pascal language took the novel approach of
providing
> neither guaranteed garbage collection nor a portable operator to free
> memory without it.  there were several platform-specific methods,
including
> stack like heap allocation (which i think was used by one compiler).
`dispose'
> ended up in the standard (but did not preclude implementing garbage
collection).

Re: [9fans] pascal, TeX

[rob pike]

> According to my knowledge - Java was the first language that implemented
> garbage collection...

Wrong.  Young people don't know any history.  The first language with
garbage collection was Visual Basic.

-rob

Re: [9fans] pascal, TeX

[Lucio De Re]

On Fri, Dec 14, 2001 at 01:03:50PM -0500, rob pike wrote:
>
> > According to my knowledge - Java was the first language that implemented
> > garbage collection...
>
> Wrong.  Young people don't know any history.  The first language with
> garbage collection was Visual Basic.
>
Rob,

this isn't alt.folklore.urban, you _are_ allowed to use smileys.

++L

Re: [9fans] pascal, TeX

[Boyd Roberts]

> The first language with garbage collection was Visual Basic.

Pity it didn't collect itself :)

IIRC VMS had an early bug (circa 1980) where the kernel managed
to swap/page the swap/page code out -- oops.

Re: [9fans] pascal, TeX

[Andrey A Mirtchovski]

<quote>
we should implement this project in VB because (as the name suggests) all
programming languages started off from basic way back when... appleII was
written in basic too!
</quote>

On Fri, 14 Dec 2001, rob pike wrote:

> > According to my knowledge - Java was the first language that implemented
> > garbage collection...
>
> Wrong.  Young people don't know any history.  The first language with
> garbage collection was Visual Basic.
>
> -rob
>
>

Re: [9fans] pascal, TeX

[Andrew Simmons]

> The first language with garbage collection was Visual Basic.
>
This still isn't quite correct. The first language with garbage collection
was Algol 68. Unfortunately, the authors of the Algol 68 revised report
referred to the process of garbage collection as "Meekly deproceduring to
MOID FORM", and nobody knew what they were talking about until Bill Gates
figured it out and told his programmers how to implement it in Visual
Basic. Bill of course went on to invent this new Internet thing everyone's
talking about.

Re: [9fans] pascal, TeX

[Thomas Bushnell, BSG]

andrew@mbmnz.co.nz (Andrew Simmons) writes:

> Bill of course went on to invent this new Internet thing everyone's
> talking about.

I thought that was Al Gore.  Oh well, live and learn.

Re: [9fans] pascal, TeX

[Douglas A. Gwyn]

rob pike wrote:
> Wrong.  Young people don't know any history.  The first language with
> garbage collection was Visual Basic.

VB certainly collects garbage, but I think he meant in a
different sense :-)

Re: [9fans] pascal, TeX

[Wladimir Mutel]

rob pike <rob@plan9.bell-labs.com> wrote:
>> According to my knowledge - Java was the first language that implemented
>> garbage collection...

> Wrong.  Young people don't know any history.  The first language with
> garbage collection was Visual Basic.

	Maybe it was LISP first ?

Re: [9fans] pascal, TeX

[Fco.J.Ballesteros]

:  :  According to my knowledge - Java was the first language that implemented
:  :  garbage collection...Neather C/C++ or Pascal (Delphi etc) have garbage
:  :  collection and memory must be manually located/delocated etc.
:  This is what's wrong with CS education today!!

Couldn't agree more!
In fact, my students know pretty well that limbo was the first
language with garbage collection.

Re: [9fans] pascal, TeX

[okamoto]

[-- Attachment #1: Type: text/plain, Size: 134 bytes --]

Oh! Please Rob, let me know it when you use some difficult rhetoric,
because there are many foreigners like me here. :-)

Kenji


[-- Attachment #2: Type: message/rfc822, Size: 2072 bytes --]

From: "Douglas A. Gwyn" <DAGwyn@null.net>
To: 9fans@cse.psu.edu
Subject: Re: [9fans] pascal, TeX
Date: Mon, 17 Dec 2001 10:14:53 GMT
Message-ID: <3C1A4FD7.8E7AB77D@null.net>

rob pike wrote:
> Wrong.  Young people don't know any history.  The first language with
> garbage collection was Visual Basic.

VB certainly collects garbage, but I think he meant in a
different sense :-)