[SCL] Re: [CL] editing questions on CL std

[Bill Andersen]

On Apr 18, 2005, at 1644, Pierluigi Miraglia wrote:

> CycL has had seqvars at least since 2001 (when I found out, and I was
> there! ;)  But they can only be in the last arg position. So one
> still can't replace all the rules that Bill intended to refer to.

Pierluigi has a good point.  There are some places you'd like seqvars 
before the last position.  I can't recall the cases right now, but we 
had some things I wanted to do with seqvars in non-final position where 
it seems that we'll need meta-CL.  But I'm not holding my breath.

Didn't know CycL had included these.  Thanks.

> One reason is that such rules are not frequently found in Cyc is that
> backchaining on unbound pred vars is (or was last I checked) a very
> dark inference sinkhole.

Yeah.  We don't use resolution-based ATP (we're using some other 
tricks) so a lot of the seqvars go away in implementation.  But they're 
nice for ontology authors.