Re: euler state of the art?

Dan Connolly wrote:
[...]
> > I tried my euler install last night, and it's lightning
> > fast, but it doesn't have all the built-ins I need.
> > In particular, I added support for fn:resolve-uri
> > and fn:tokenize from XQuery to cwm this week.

> It is pretty straightforward to write builtins in euler.yap
> http://eulersharp.sourceforge.net/2006/02swap/euler.yap
> but I should study your builtins a little more..

I haven't made much progress yet with some of those builtins..
What I could do was http://eulersharp.sourceforge.net/2006/02swap/DONE
i.e.

- adding list:in, list:member and log:uri builtins
- correction in proof output; thanks to Dan Connolly
- add http port number support

and the latest running code is at
http://eulersharp.sourceforge.net/2004/01swap/deployment/Euler.zip

The list:in we had before was done via
http://eulersharp.sourceforge.net/2003/03swap/rpo-rules.n3
but it is now a builtin and makes proofs like
http://eulersharp.sourceforge.net/2006/02swap/gedcomE.n3
also a bit smaller and check.py is still happy with the proof output
i.e.

[...]
643: :Dirk :testRelationship :Bart, :Tom . :Geert :testRelationship :Bart, 
:Tom . :Jos :testRelationship :Bart, :Tom .
 [by CI on [98, 196, 300, 404, 518, 642]]


I was not sure about list:member, but implemented it as inverse of list:in
not sure because the python doc of 
http://www.w3.org/2000/10/swap/cwm_list.py
says

class BI_member(LightBuiltIn, MultipleFunction):
    """Is the subject in the object?
    Returnes a sequence of values."""

which is similar to

class BI_in(LightBuiltIn, MultipleReverseFunction):
    """Is the subject in the object?
    Returnes a sequence of values."""

Anyhow, to test builtins, I use a growing bunch of test cases in
http://www.agfa.com/w3c/euler/testbuiltins.n3
giving proof output
http://eulersharp.sourceforge.net/2006/02swap/testbE.n3

More to be done..


-- 
Jos De Roo, AGFA http://www.agfa.com/w3c/jdroo/