Re: other formalisms

The 2019 W3 Graph Workshop in Berlin searched for ways to bridge the gap between RDF and LPG and improve interoperability. Statement annotation was obviously the most pressing missing piece in RDF, given that standard reification is so verbose and NG have no semantics. Olaf successfully positioned RDF* as an approach that could help in this respect. A CG was formed to work out the details. A model-theoretic semantics  had to be developed (Olaf had advertised RDF* to the WS  as having a semantics, but well...).  Now a WG has been formed to continue the work of the CG. My understanding is that it is - still - the foremost goal of the WG to standardize a concise and sound statement annotation mechanism that allows to express LPG annotated statements in RDF.

Now as we all know there is no standard definition of LPG, but the gist is pretty straightforward: statements can be annotated with just about any kind of information, especially NOT ONLY PROVENANCE. Any kind of pseudo "prudent" approach that discourages or warns about or just not mentions anything but provenance use cases (as is the approach taken by the CG  report) will be ignored in practice. If the syntax is only halfway attractive it will be used for just about anything. If the semantics can't handle that it will be useless, or harmful even. 

I have no idea what you refer to when you say "and databases" - maybe something from the meeting that wasn't recorded in the minutes? Or do you mean my mail about implementation issues?

Best,
Thomas 

Am 23. Januar 2023 19:05:33 MEZ schrieb "Peter F. Patel-Schneider" <pfpschneider@gmail.com>:
>Several other formalisms have shown up in the working group emails, such as property graphs and databases.  It is hard to refute claims about a formalism when there is no definition and implementation to evaluate the claims in.  This can easily lead to wastage of WG meeting time.
>
>So please, if you want to make a claim about some formalism please provide a definition for that formalism and point to an implementation of it.  I think that if this is not done then any claims should only be allowed if they are true in all known related formalisms and implementations.
>
>
>peter
>
>
>