Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.programming > #16925 > unrolled thread

continuations in ``the art of computer programming''

Back to article view | Back to comp.programming

Contents

  continuations in ``the art of computer programming'' Anton Antimo <anton@safunu.org> - 2026-09-22 13:43 -0300
    Re: continuations in ``the art of computer programming'' ram@zedat.fu-berlin.de (Stefan Ram) - 2026-09-22 18:04 +0000
    Re: continuations in ``the art of computer programming'' usenet@stegropa.de (Stefan Große Pawig) - 2026-09-22 20:01 +0200
      Re: continuations in ``the art of computer programming'' Lawrence D’Oliveiro <ldo@nz.invalid> - 2026-09-22 21:07 +0000
      Re: continuations in ``the art of computer programming'' Anton Antimo <anton@safunu.org> - 2026-09-24 13:59 -0300
        Re: continuations in ``the art of computer programming'' Lawrence D’Oliveiro <ldo@nz.invalid> - 2026-09-25 00:06 +0000

#16925 — continuations in ``the art of computer programming''

Donald Knuth has been documenting and investigating programming
techniques forever in his monumental set called ``The Art of Computer
Programming''.  I suppose he must have in his own terms what is now
called continuations.  In which part section of which book is the Scheme
call/cc primitive studied?  

[toc] | [next] | [standalone]

#16926

Anton Antimo <anton@safunu.org> wrote or quoted:
>Donald Knuth has been documenting and investigating programming
>techniques forever in his monumental set called ``The Art of Computer
>Programming''.  I suppose he must have in his own terms what is now
>called continuations.  In which part section of which book is the Scheme
>call/cc primitive studied?  

  Scheme's call/cc is a high-level language programming con-
  struct outside the scope of Knuth's assembly-based approach.
  However, coroutines are treated in Volume 1, section 1.4.2.

  "call/cc" could be treated in volume 7 "Compiler Techniques",
  but volume 7 is still decades away.

[toc] | [prev] | [next] | [standalone]

#16927

Anton Antimo <anton@safunu.org> writes:
> Donald Knuth has been documenting and investigating programming
> techniques forever in his monumental set called ``The Art of Computer
> Programming''.  I suppose he must have in his own terms what is now
> called continuations.  In which part section of which book is the Scheme
> call/cc primitive studied?  

I think the closest would be volume 1, section 1.4.2 on coroutines.

  Regards,
    Stefan

[toc] | [prev] | [next] | [standalone]

#16928

On Tue, 22 Sep 2026 20:01:18 +0200, Stefan Große Pawig wrote:

> Anton Antimo <anton@safunu.org> writes:
>>
>> In which part section of which book is the Scheme call/cc primitive
>> studied?
>
> I think the closest would be volume 1, section 1.4.2 on coroutines.

Note that (stackful) coroutines are just one of the higher-level
constructs that can be implemented with a continuation primitive.

Continuations are somewhat more general than that.

[toc] | [prev] | [next] | [standalone]

#16929

usenet@stegropa.de (Stefan Große Pawig) writes:

> Anton Antimo <anton@safunu.org> writes:
>> Donald Knuth has been documenting and investigating programming
>> techniques forever in his monumental set called ``The Art of Computer
>> Programming''.  I suppose he must have in his own terms what is now
>> called continuations.  In which part section of which book is the Scheme
>> call/cc primitive studied?  
>
> I think the closest would be volume 1, section 1.4.2 on coroutines.

Thank you and all other posters in this thread.

So I think call/cc is essentially a jump over there, do something and
take here an address for you to get back to what you were doing before.

As usual, not much magic behinds the scenes, but surely the implications
of a jump instruction as far fetching, of which call/cc is a great
illustration.

[toc] | [prev] | [next] | [standalone]

#16930

On Thu, 24 Sep 2026 13:59:54 -0300, Anton Antimo wrote:

> So I think call/cc is essentially a jump over there, do something
> and take here an address for you to get back to what you were doing
> before.

There doesn’t need to be a jump to begin with. You could just save the
return point and continue execution normally. Think setjmp/longjmp in
C -- that’s a form of continuation, too.

It’s just that the Lisps usually implement it as “call a function
while passing it another function which, when called, jumps back to
just after the point of the initial call”.

[toc] | [prev] | [standalone]

Back to top | Article view | comp.programming

csiph-web