W3C home > Mailing lists > Public > public-rif-wg@w3.org > November 2007

Re: (ISSUE-40) Builtins and logic functions in BLD

From: Michael Kifer <kifer@cs.sunysb.edu>
Date: Tue, 20 Nov 2007 21:01:47 -0500
To: Christian de Sainte Marie <csma@ilog.fr>
Cc: Dave Reynolds <der@hplb.hpl.hp.com>, public-rif-wg@w3.org
Message-ID: <5823.1195610507@cs.sunysb.edu>

> Michael Kifer wrote:
> >>Michael Kifer wrote:
> >>
> >>>I made a proposal that we should treat builtins using the same mechanism as
> >>>modules. For instance, if a builtin is defined in the XQuery/XPath library
> >>>then we would refer to it as
> >>>
> >>>    fn:dateTime(...)@http://www.w3.org/2005/xpath-functions
> >>>
> >>>where fn is a prefix for http://www.w3.org/2005/xpath-functions
> You mean that "http://www.w3.org/2005/xpath-functions/#dateTime" 
> (possibly abbreviated "fn:dateTime") is a Const of type rif:iri, right?
> Then, I agree with Dave:
> >Dave Raynolds wrote:
> >>Isn't the URI enough to avoid clashes?
> Why do you need more than that to identify built-ins?

I do not. (see the quoted text below)

> > For builtins the module system is not needed. It was just one way to
> > indicate that we are dealing with something that is defined by an external
> > library.  I recall that people did not like the idea of deciding whether
> > something is a builtin or not based solely on iris.
> Could those people (who did not like the idea of deciding wether...) 
> explain why (they did not like that idea) to me?
> Also, are we talking about external calls (prcedural attachments) in RIF 
> in general, or about RIF-BLD built-ins, here?
> If we are talking about RIF-BLD built-ins, I do not understand the 
> discussion about having to decide whether someting is a built-in or not: 
> aren't built-ins listed and specified as part of the RIF-BLD 
> specification? It seems to me that here is nothing to decide: either 
> "http://www.w3.org/2005/xpath-functions/#dateTime" is listed in the 
> RIF-BLD spec as a built-in, or it is not. Did I miss something?

Just like we want people to make rif datatype-extensible, we should allow
partners to exchange rulesets that use their private libraries of builtins
(which both partners know about).

> > But, on the other hand, the same builtin may be defined by different
> > libraries, and the module system may open a way to use different libraries.
> Are you talking about different implementations of the same built-in? 
> Here again, if we are talking about RIF-BLD built-ins, isn't that out of 
> scope?

Why is it out of scope? This kind of considerations are a fair game.
I am not saying that this is what I would push, but this kind of
extensibility is not a bad idea.


> Christian
Received on Wednesday, 21 November 2007 02:02:04 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:47:48 UTC