W3C home > Mailing lists > Public > semantic-web@w3.org > June 2020

Re: defining the semantics of lists

From: Thomas Passin <tpassin@tompassin.net>
Date: Wed, 3 Jun 2020 18:24:32 -0400
To: Hugh Glaser <hugh@glasers.org>, Peter Patel-Schneider <pfpschneider@gmail.com>
Cc: thomas lörtsch <tl@rat.io>, semantic-web@w3.org
Message-ID: <1e0ba3d6-61ec-ff19-1124-c6fb7e59f8f2@tompassin.net>
On 6/3/2020 6:02 PM, Hugh Glaser wrote:
> +1
> I've followed the long discussion (as best I could!).
> And what I kept wanting to know was an example of how whatever is being proposed might be used.
> I remember someone (Pat?) rightly saying that RDF was not a data structure language, but a KR (or something like that).
> So simply wanting the sort of things that programming languages have as data structures is not necessarily a useful thing to spend time on.


I still claim that many if not most lists are ephemeral, and therefore 
are not good candidates for inclusion in an RDF graph.  They are 
ephemeral because they are subject to re-sorting into a different order 
at any time.  So they are not stable facts about the world, or some 
worlds.  What interpretations of a graph would change if the list were 
to be resorted?  If none, then why pick one particular sort order, and 
therefore why have a list?  If otherwise, it's hard to say anything 
general about the changed set of interpretations.

Now if the elements of what would otherwise be a list had properties 
that can be sorted on, then they could be sorted into lists by software 
outside of the RDF machinery.  This is similar to the situation in SQL 
databases, which do not know about the order of rows.  The query 
processor may be asked to to sorting, but it's not inherent in the 
stored data.

It works for SQL databases.  Why not for RDF knowledge sets?  Remember, 
if you normalize an SQL database down to tables with just two columns, 
you have a very close equivalent to a collection of RDF triples - not 
counting the nature of the identifiers, of course.  By looking at it 
this way, it seems that SPARKL should do any sorting needed.

And yes, people do represent lists in SQL.  And look how clumsy that 
turns out to be.
Received on Wednesday, 3 June 2020 22:24:53 UTC

This archive was generated by hypermail 2.4.0 : Tuesday, 5 July 2022 08:46:04 UTC