Re: Why semantic markup on public web pages in 2017

Hi Chris,

I see two questions in your mail:
1. Why semantic markup on webpages?
2. Why use the Dublin Core vocabulary for that?

The first question is something I've been wondering for some time,
as I have put quite some effort through the years in writing RDFa
on my own website, without knowing whether it would make any impact.

I recently found some new motivation
by making all of that data queryable in a single interface:
– data: https://data.verborgh.org/ruben
– queries: http://query.verborgh.org/
This exposes all of my RDFa and derived triples,
and people can ask targeted questions over that data.

This might seem a counterexample,
as having a single interface could void the need for RDFa.
However, I still see several benefits:
– For me as an author, it is easier to annotate on individual pages
   than having to maintain a collection of triples (how?).
– Search engines get access to the markup they need
   on the right page.
– Browser extensions can exploit a page's RDFa.

I discuss these matters at length in an upcoming LDOW2017 submission:
https://ruben.verborgh.org/articles/queryable-research-data/
(this text is not final yet, but comments are welcome).

As for your second question, my answer is simple:
I think the DC vocabulary is too weak for most practical purposes.
This is why my webpages mostly use Schema.org
(which is weak in other ways).
However, relevant DC properties are automatically derived from it
by the reasoner that is behind my queryable interface;
yet these properties are not on the individual pages.

Best,

Ruben

Received on Thursday, 19 January 2017 09:35:02 UTC