jonny (good kind)<p><span class="h-card" translate="no"><a href="https://hachyderm.io/@hrefna" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>hrefna</span></a></span> you're so deeply right in this thread. this essentially defines the early and formative years of <a href="https://neuromatch.social/tags/RDF" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RDF</span></a> and <a href="https://neuromatch.social/tags/RDFS" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RDFS</span></a> around some key decisions. There is <em>so much worry</em> about errors from mishandling, but the <em>very possibility</em> of expressing something that <em>could</em> be wrong. There is so much here it's hard to even know where to start but from the very start of the rdf-core and especially rdf-logic mailing lists the emphasis on correctness of expression vs. agency of interpretation is heated. This is a pretty unsubtle example: <a href="https://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Nov/0423.html" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">lists.w3.org/Archives/Public/w</span><span class="invisible">3c-rdfcore-wg/2001Nov/0423.html</span></a></p><blockquote><p>"Summary: RDF is not just a data model. The RDF specs should define a semantics so that an RDF statement on the web is interpreted as an assertion of that statement such that its author would be responsible in law as if it had been published in, say, a newspaper."</p></blockquote><p>affirmed by this text: <a href="https://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Jun/0180.html" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">lists.w3.org/Archives/Public/w</span><span class="invisible">3c-rdfcore-wg/2002Jun/0180.html</span></a></p><blockquote><p>"The RDFCore WG takes the view that RDF/XML documents, ie. encodings of<br>RDF graphs, are represenations of claims or assertions about the world. [...] we note that RDF graphs are the kind of things that can be true or false (rather than 'mere bytes'). For each linear syntax of RDF, there will be conventions (social, legal) for indicating which chunks of data are encodings of RDF graphs in that syntax, and thus of propositional content."</p></blockquote><p>So ya very very early they had decided that RDF was intended to be platonic representation of the world - it is a method of making propositions that must directly bear on the world, there is no ambiguity, there are only true or false things. (There's a lot of subtlety in that but yno.)</p><p>Another thing that comes to mind is the discussion of monotonicity that happens throughout, that's also tied up with the discussion of closed worldedness, containers, and whether or not things belong in sets for the purpose of local meaning or need to belong to the One Great Uniform Graph. It must be impossible for adding new triples into the reasoner to make it less certain of some outcome (ie. if it was possible for ambiguity to be introduced eg. by tangled type hierarchies etc.). This thread is all very illustrative but this message is super spicy and also touches on what you're talking about re using the level of granularity you care about: <a href="https://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Nov/0363.html" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">lists.w3.org/Archives/Public/w</span><span class="invisible">3c-rdfcore-wg/2001Nov/0363.html</span></a></p><p><a href="https://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Nov/0381.html" rel="nofollow noopener" target="_blank">this quote</a> from D bricks is also pretty telling of the epistemological commitments: <br>"RDF Schema is not about policing the content of specific descriptions<br>'what is said'; it's about policing what is sensibly sayable."</p>