W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > June 2001

Re: try 2 - test cases for #rdf-containers-syntax-ambiguity, #rdf-containers-syntax-vs-schema

From: Brian McBride <bwm@hplb.hpl.hp.com>
Date: Fri, 15 Jun 2001 00:40:02 +0100
Message-ID: <3B294B52.E64E386@hplb.hpl.hp.com>
To: Aaron Swartz <aswartz@upclink.com>
CC: RDF Core <w3c-rdfcore-wg@w3.org>


Aaron Swartz wrote:
[...]
> Can you please run these tests through SiRPAC/CWM before
> submitting them next time. That should catch these silly typos.

Absolutely right.  But thanks anyway for picking them up.

> 
> >         Whether this represents a legal collection of triples
> >         depends on the definition of foo:Bar.  If foo:Bar is
> >         subclass of container then it is illegal.
> 
> Can you elaborate on why this is? I believe it is because of the
> functional container properties thing, which we as a working
> group have not yet made a decision on to my knowledge (do they
> only apply to containers or any resource?).

What decision do you think the WG has to make?  M&S says:

  For a single collection resource there may be at most one triple
  whose predicate is any given element of Ord.

Members of Ord are not constrained to only apply to containers.
There is no cardinality restriction stated for Ord in general.
No issue has been raised proposing such a constraint. 

> 
> >         rdf:li elements as typed nodes - a bizarre case
> >         but handled regularly
> 
> I'm not sure that this is specified in the M&S (perhaps this
> should be an issue) but it's certainly not a good idea to make
> something both a class and a property. Let's not have a test
> case that does this, please. 

I suggest that a test case is need for this so that it is
clear how an <rdf:li> as a typed node should be processed.  Is it
an error, or should it be translated to <rdf:_nnn>, or does it
define a node with type rdf:li?

I'm open minded about whether this should be an error or should
be legal but strange.  I had two reasons for suggesting that it
should be legal and processed as I suggested:

  o a general principle that one should not rule things illegal
    just because they look a little strange - someone might come
    up with an unforseen use for it.

  o special cases are best avoided.  As this case seems to do no
    harm then why go to the trouble of outlawing it.  This way
    the XSLT to transform <rdf:li> to rdf:_nnn is more
    straightforward, for example.

However, I do seek the WG's advice on how this might best be handled.

> Especially since rdf:li is a
> special syntactic element -- I do not think we should allow it
> to be used in this way.

I don't follow your reasoning there, but +1 for its an error.

Brian
Received on Thursday, 14 June 2001 19:43:07 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:37:08 EDT