Do parsers apply the value of "base" elements when resolving URLs?

[w3c]

Do parsers apply the value of "base" elements when resolving URLs?

[iherman]

@gkellogg knows me well, but, to be on the safe side, I checked my code too:-) Yes, RDFLib & Co. do implement the HTML <base> element.

[gkellogg]

RFC3986 has a section on URI resolution, and both HTML and XML have specific rules for resolving relative URIs against an established base.

For HTML, it's typically the value of the base element in the head section. XML uses xml:base, on the closest element; many RDFa parsers do this as well.

In the absence of an explicit base, it is set from the document location.

As RDF does not allow relative IRIs in the abstract syntax, everything needs to be resolved against something. example.org has no special meaning other than convention.

[chaals]

Yeah, I am specifically wondering if microdata parsers actually do it...

[gkellogg]

With regards to RDF Microdata parsers, there are tests to ensure this is how relative URIs are handled. Mine does this, and I'm sure @iherman's does as well.

There are so many variations, @RubenVerborgh wrote a number of tests for various RDF serlializations to test this. We repeated those for JSON-LD, but haven't gone that far in the RDF or Microdata RDF test suites.

[chaals]

This is less important than I thought, if we solve #41 so typed items collect the vocabulary we expect for describing the names of URLs.

[chaals]

In any event, I deferred to RFC3986 for resolving URLs, so I think that covers it and we can close this issue.

@gkellogg ?

[chaals]

I did a quick'n'dirty test on the Google validator and it applied base too. So nothing to see here...