No subject

[jc at minster.cs.york.ac.uk]

>From jc Mon Jun  8 11:35 BST 1998
From: James Cussens <jc at cs.york.ac.uk>
To: mercury-users at cs.mu.OZ.AU
Subject: partially instantiated data strucutures
Content-Type: text
Content-Length: 1539


(1) 
According to the Modes node of the information documentation: 

Nevertheless, Mercury's mode system is sufficiently expressive to
handle more complex data-flow patterns, including those involving
partially instantiated data structures.  (The current implementation
does not handle partially instantiated data structures yet.)

Are there any plans to allow partially instantiated data structures to
be handled. If there is some reasonable way of handling them at
present, then I've missed it.

(2) 
I have a couple of students doing a final year project on "Using
Mercury for Inductive Logic Programming". This will start in the
autumn. An ILP algorithm constructs a logic program (H) from examples
(E) and background knowledge (B) which are also logic programs. In the
simplest case, E is split into positives and negatives; H,B should
entail the positives and not the negatives (and not be too "complex",
otherwise you could just have H as the positives).

The motivation for using Mercury is 
a) You generally have to declare types, modes etc so your algorithm
knows how to find H.
b) Speed

Mercury seems entirely suitable as a language for an ILP algorithm to
me, but if anyone can see any likely problems or knows of any future
developments of Mercury that are relevant I would like to know.

James

-- 
James Cussens 				             jc at cs.york.ac.uk
Department of Computer Science 		      Tel  +44 (0)1904 434732 
University of York			      Fax  +44 (0)1904 432767
Heslington, York YO10 5DD, UK		 http://www.cs.york.ac.uk/~jc