;
To: "Claus Reinke" <[EMAIL PROTECTED]>
Cc:
Sent: Monday, March 13, 2006 10:57 PM
Subject: Re: alternative translation of type classes to CHR(was:relaxedinstance
rules spec)
On 3/13/06, Claus Reinke <[EMAIL PROTECTED]> wrote:
[still talking to myself..?]
This is all won
On 3/13/06, Claus Reinke <[EMAIL PROTECTED]> wrote:
> [still talking to myself..?]
This is all wonderful stuff! Are you perhaps planning to put it all
together into a paper?
What effect do you think this can have on existing algorithms to resolve FDs?
--
Taral <[EMAIL PROTECTED]>
"Computer scien
[still talking to myself..?]
all confluence problems in the FD-CHR paper, as far as they were
not due to instances inconsistent with the FDs, seem to be due to
conflicts between improvement and inference rules. we restore
confluence by splitting these two constraint roles, letting inference
a
a second oversight, in variation B: CHR rules are selected by matching,
not by unification (which is quite essential to modelling the way type
class inference works). this means that the idea of generating memo_
constraints for the instance fdis and relying on the clas fdi rules to
use that info