quintuples and Re: "whole reason quads were implemented"

On 5/10/2012 1:12 AM, Dan Brickley wrote:
>>> Which systems?  I'd like to understand the motivations and approaches.
>> I don't actually think the history is that useful.  I was just clumsily
>> trying to get Steve to stop generalizing and consider some of the other
>> perspectives or help me understand why he wouldn't/couldn't.
>> But since you ask, and now I'm thinking about it....

The Named Graphs paper was trying to be a thinking person's summary of the state of the art in the early phases ... one of the key inputs was cwm and N3, where logic and inference were key motivators. Personally I didn't much like CWM approach to logic, but the use case was certainly one that was motivating, as were the provenance use cases that got into the title.

Notice that 'trust' requires a mix of provenance and reasoning.

So, I'm with Sandro, it is probably unhelpful to focus in on only our own experience, and the point of a WG is to bring together many people's experiences.

A further insight from early on, that is relevant, was a discussion me and Chris Bizer had while he was an intern at HP labs, in the early stages of the research for that paper.

He initially had a system (on paper) with quintuples as opposed to quads or triples. We did some thought experiments, and showed that the use cases that appeared to require quintuples, could in fact be addressed with quads, while triples (and reification) seem to consistently fail to address the use cases satisfactorily. So from my perspective quads got implemented because:
- they addressed several use cases significantly better than triples and reification
- quintuples or n-tuples did not offer big enough advantages over quads

Being a verbal discussion there aren't any records and my memory fades - once Patrick and Pat got in on it, we started archiving e-mail discussion on www-archive


Received on Thursday, 10 May 2012 17:30:32 UTC