- From: Henry Story <henry.story@bblfish.net>
- Date: Thu, 16 Jul 2020 16:59:18 +0200
- To: David Booth <david@dbooth.org>
- Cc: semantic-web <semantic-web@w3.org>, Evan Patterson <epatters@stanford.edu>, dspivak@math.mit.edu
> On 16 Jul 2020, at 16:18, David Booth <david@dbooth.org> wrote: > > On 7/16/20 9:58 AM, Dan Brickley wrote: >> I believe the big appeal of putting it all into the zone we call "literals" is that you get a kind of atomicity; that chunk of data is either there, or not there; it is asserted, or not asserted. With a triples-based (description of a ) data structure you have to be constantly on your guard that every subset of the full graph pattern is at least sensible and harmless, even when subsetting these chunks is often confusing or misleading for data consumers. I can't help wondering whether notions of graph shapes [ . . . ] could be exploited to create an RDF-based data format which had atomicity at the level of entire shapes. > > +1 > > IMO the ability to manipulate chunks of data atomically -- arrays, n-ary tuples and hierarchical objects -- is a key requirement in developing a higher-level form of RDF. This will include the need to conveniently construct and deconstruct such chunks in rules or query languages. > > David Booth I think we are getting somewhere close to the idea of typed RDF. I don’t yet have opinions on that yet, but have come across it reading a paper by Even Patterson ”Knowledge Representation in Bicategories of Relations" that extends the ideas developed in Functorial Databases developed at MIT by David Spivak to the category Rel of Sets and Relations (which intuitively fits RDF better, though there is a translation between the two). It turns out that the Grothendieck Construction of a functor from a small bicateogry of Relations to Rel, gives a typed RDF. See: https://gitlab.com/web-cats/CG/-/issues/9 Is RDF typed? Well it may be that it is partly, or that it could be moved in that direction. OWL partly goes in that direction with ObjectProperties and DataProperties. If one thinks of objects as forming one big type not containing literals, and each literal type forming its own space, then one has have a light weight type system in RDF. This may have reasoning advantages. (mathematicians would need to be consulted). For example it looks like OWL allows one to specify a subtype of xsd:int that does not contain a certain range of numbers. Does it really make sense to think of things that could satisfy those as containing elephants too? Would it not make more sense to understand those as still covering integers? I think so. And it is quite easy to see that this would make reasoning easier, as the space to search would be smaller). It is also interesting to note on this topic that Functorial databases that have a beautiful simiplicity, had to be enhanced around 2013 by adding a construction to fix types. Without that the morphisms between database instances (functors) would allow names of people to also be exchanged when the structures were isomorphic. Essentially all elements of the database including literals were represented as blank nodes. This points to the distinction between literals and objects as being fundamental to all databases. I am not sure if Evans has integrated the work on functorial databases that keeps literals constant to his bicategories of relations. It would probably be easy to translate the findings as there is an isomorphism I understand between the two structures. Henry Story https://co-operating.systems WhatsApp, Signal, Tel: +33 6 38 32 69 84 Twitter: @bblfish
Received on Thursday, 16 July 2020 14:59:33 UTC