RE: Proposed issue resolutions

At 02:11 PM 11/7/01 +0000, Jeremy Carroll wrote:



>I had assumed that:
>
>    aaa <rdf:type> <rdf:Bag> .
>
>    aaa <rdf:_NNN> bbb .
>    aaa <rdf:_MMM> ccc .
>
>
>
>=>
>
>   aaa <rdf:_MMM> bbb .
>
>
>where aaa is a bNode or URI and bbb, ccc are anything.
>
>(This means that a bag with n members licenses n^2 edges ).

I don't think that works.  Consider a bag {* "aaa" "bbb" "bbb" *}

The 3^2=9 edges would be:

    aaa rdf:_1 "aaa" .
    aaa rdf:_1 "bbb" .
    aaa rdf:_1 "bbb" .
    aaa rdf:_2 "aaa" .
    aaa rdf:_2 "bbb" .
    aaa rdf:_2 "bbb" .
    aaa rdf:_3 "aaa" .
    aaa rdf:_3 "bbb" .
    aaa rdf:_3 "bbb" .

But, on removing duplicates, which the MT tells us are redundant:

    aaa rdf:_1 "aaa" .
    aaa rdf:_1 "bbb" .
    aaa rdf:_2 "aaa" .
    aaa rdf:_2 "bbb" .
    aaa rdf:_3 "aaa" .
    aaa rdf:_3 "bbb" .

Now does that bag contain 2 "aaa"s and 1 "bbb", or 1 "aaa" and 2 "bbb"s, or 
something else?

#g


------------------------------------------------------------
Graham Klyne                    MIMEsweeper Group
Strategic Research              <http://www.mimesweeper.com>
<Graham.Klyne@MIMEsweeper.com>
------------------------------------------------------------

Received on Wednesday, 7 November 2001 16:13:06 UTC