W3C home > Mailing lists > Public > public-rdf-prov@w3.org > October 2011

Re: complete graphs

From: William Waites <ww@styx.org>
Date: Tue, 04 Oct 2011 13:25:36 +0200
To: Andy Seaborne <andy.seaborne@epimorphics.com>
Cc: public-rdf-prov@w3.org
Message-ID: <86ipo52e33.fsf@river.styx.org>
>>>>> "andy" == Andy Seaborne <andy.seaborne@epimorphics.com> writes:

    >> If we interpreted the fourth column in this way, what would
    >> happen?

    andy> Reification.

Not sure this is necessarily the case. If you were to use assertion
identifiers to construct graphs it would look like reification but this
is also an obvious candidate for a specific optimisation in storage
systems. Apart from that it would mean we can talk about the provenance
of statements as well as (or instead of) the provenance of collections
of statements.

Why might this be useful? If you accept that data is generally dirty and
that trying to do inference on it using only formal logic is
unreasonably idealistic, you can start doing things like saying,

    I have assertion A from two sources S1 and S2. Either of those
    sources on their own aren't particularly trustworthy, but taken
    together, I decide to believe A with a confidence of P.

I'm not sure there's a straightforward way to express this right now
short of minting lots of sub-graphs of size 1. These sub-graph
identifiers start looking like statement identifiers and the X
subGraphOf Y statements start looking the same as what I believe you are
pointing to as the reification...

-w

-- 
William Waites                <mailto:ww@styx.org>
http://river.styx.org/ww/        <sip:ww@styx.org>
F4B3 39BF E775 CF42 0BAB  3DF0 BE40 A6DF B06F FD45
Received on Tuesday, 4 October 2011 11:26:07 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 4 October 2011 11:26:08 GMT