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

Groups > comp.lang.java.programmer > #53775 > unrolled thread

Dogelog in the Black Forest: The Ouroboros Ring Road Puzzle (Was: ANN: Dogelog Player 2.3.1 (Task Channels))

Back to article view | Back to comp.lang.java.programmer

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Dogelog in the Black Forest: The Ouroboros Ring Road Puzzle (Was: ANN: Dogelog Player 2.3.1 (Task Channels)) Mild Shock <janburse@fastmail.fm> - 2026-09-20 17:43 +0200
    Dogelog: Getting the birds from the Byrd Box model (Was: Dogelog in the Black Forest: The Ouroboros Ring Road Puzzle) Mild Shock <janburse@fastmail.fm> - 2026-09-27 15:40 +0200

#53775 — Dogelog in the Black Forest: The Ouroboros Ring Road Puzzle (Was: ANN: Dogelog Player 2.3.1 (Task Channels))

Hi,

In this post, we look back at working with cyclic
terms in a Prolog system, a feature that was
introduced in Dogelog Player around 12 months ago.
We welcome current developments like SWI-Prologs
recent term_minimum/2. They serve us as templates
to name our own library predicates.

While coinductive predicates are still in its
infancy for Prolog systems. We could nevertheless
solve the Ouroboros Ring Road puzzle via a cyclic
term modelling. The article also touches differences
among Prolog systems in cyclic term display and
the bagof/3 use case of ordering cyclic terms.

Bye

See also:

Dogelog in the Black Forest: The Ouroboros Ring Road Puzzle
https://medium.com/2989/147fc5648b72

Mild Shock schrieb:
> Dear All,
> 
> We are happy to announce a new
>   edition of the Dogelog Player:
> 
> - Enhanced catch/3:
> We noticed that catch/3 causes a friction in
> '$YIELD'/1. So we redesigned its implementation
> by refraining from using the host programming
> language stack. The engine now offers an exception
> handler stack, and the Prolog interpreter trampolin
> unwinds these exception handlers upon a
> Prolog or host error.
> 
> - New call_cleanup/3:
> The new predicate call_cleanup/3 supplants the
> old predicate once_cleanup/3. Like catch/3 it
> now features an implementation that refrains
> from using the host programming language stack.
> Its implementation makes use of a rogue '$CUT'/1
> which is realized by choicepoint scanning
> and continuation manipulation.
> 
> - Task Channels:
> To lower the footprint of the core there are
> new libraries library(timer) and library(tasks).
> The later library offers a new predicate task_join/1
> which is abortable. Further the library offers
> predicates chan/1, recv/2 and send/2 to deal with
> channels that live in the cooperative
> multitasking realm.
> 
> Have Fun!
> 
> Jan Burse, September 17, 2026, https://www.herbrand.ai/

[toc] | [next] | [standalone]

#53776 — Dogelog: Getting the birds from the Byrd Box model (Was: Dogelog in the Black Forest: The Ouroboros Ring Road Puzzle)

Hi,

It turns out some things are more complicated than
the mind body problem. Without seeing an anlogue
behind our recent rogue ‘$CUT’/1 and the requirements
of an undo handler. It would have taken us another
couple of years before implementing the
said gadget.

The new addition is a refined class hierarchy, where
a class Store.Bind forms the base class of both
Store.Variable and Store.Undo. While we can demonstrate
the classical set_arg/3, a backtrackable change_arg/3,
we have doubts whether we are already on the
right path forward.

Bye

See also:

Dogelog: Getting the birds from the Byrd Box model
https://medium.com/2989/e75796eaa2df

Mild Shock schrieb:
> Hi,
> 
> In this post, we look back at working with cyclic
> terms in a Prolog system, a feature that was
> introduced in Dogelog Player around 12 months ago.
> We welcome current developments like SWI-Prologs
> recent term_minimum/2. They serve us as templates
> to name our own library predicates.
> 
> While coinductive predicates are still in its
> infancy for Prolog systems. We could nevertheless
> solve the Ouroboros Ring Road puzzle via a cyclic
> term modelling. The article also touches differences
> among Prolog systems in cyclic term display and
> the bagof/3 use case of ordering cyclic terms.
> 
> Bye
> 
> See also:
> 
> Dogelog in the Black Forest: The Ouroboros Ring Road Puzzle
> https://medium.com/2989/147fc5648b72
> 
> Mild Shock schrieb:
>> Dear All,
>>
>> We are happy to announce a new
>>   edition of the Dogelog Player:
>>
>> - Enhanced catch/3:
>> We noticed that catch/3 causes a friction in
>> '$YIELD'/1. So we redesigned its implementation
>> by refraining from using the host programming
>> language stack. The engine now offers an exception
>> handler stack, and the Prolog interpreter trampolin
>> unwinds these exception handlers upon a
>> Prolog or host error.
>>
>> - New call_cleanup/3:
>> The new predicate call_cleanup/3 supplants the
>> old predicate once_cleanup/3. Like catch/3 it
>> now features an implementation that refrains
>> from using the host programming language stack.
>> Its implementation makes use of a rogue '$CUT'/1
>> which is realized by choicepoint scanning
>> and continuation manipulation.
>>
>> - Task Channels:
>> To lower the footprint of the core there are
>> new libraries library(timer) and library(tasks).
>> The later library offers a new predicate task_join/1
>> which is abortable. Further the library offers
>> predicates chan/1, recv/2 and send/2 to deal with
>> channels that live in the cooperative
>> multitasking realm.
>>
>> Have Fun!
>>
>> Jan Burse, September 17, 2026, https://www.herbrand.ai/
> 

[toc] | [prev] | [standalone]

Back to top | Article view | comp.lang.java.programmer

csiph-web