- From: William Waites <wwaites@tardis.ed.ac.uk>
- Date: Fri, 23 Nov 2018 13:26:34 +0000
- To: Hugh Glaser <hugh@glasers.org>
- Cc: Wouter Beek <wouter@triply.cc>, tpassin@tompassin.net, SW-forum Web <semantic-web@w3.org>
> So my tools are akin to Assemblers or Compilers for primitive > languages… So, if we achieve an easier RDF stack, then Blank > Nodes become a complete non-issue, just like registers are in > High-Level languages. This is exactly right with one small change. If you think of RDF as simply data then yes, you are right, you’ve got all sorts of high-level syntaxes that you can get triple “machine code” from. It is the input to (and possibly the output from) a program that does a useful computation. In practice we almost always translate the uniform triples into some sort of specialised data structure internal to the program. Working with triples directly is verbose and unwieldy so we query them and put the result into some data structure that is more convenient and work with that. We can also think of it, with appropriate vocabularies, as a way to write programs in a declarative style. Then the triples are a kind of intermediate representation that needs to be translated into an executable program. This hearkens back to the “code as data” idea from LISP which is where, if I understand the history correctly, some of the ideas for RDF derive from. Assemblers and compilers and program generators are very useful ways to think about this stuff. Cheers, -w
Received on Friday, 23 November 2018 13:27:01 UTC