Hacker News Reader

Comment by est | Hacker News Reader

I've playing with RSS feeds recently, suddently it occured to me, XML can be transformed into anything with XSL, for static hosting personal blogs, I can save articles into the feeds directly, then serve frontend single-page application with some static XSLT+js. This is content-presentation separation at best.

Is JSON-LD just reinventation of this?

View on HN

Back in the optimistic 2000s there was the brief idea of GRDDL – using XSLT stylesheets and XPath selectors for extracting stuff from HTML, e.g. microformats, HTML meta, FOAF, etc, and then transforming it into RDF or other things:

https://www.w3.org/TR/grddl/

But why? Isn't most of the information you can extract from those tags stuff that's pretty obvious, like title and author (the examples the linked page uses)? How do you extract really useful information using that methodology, supporting searches that answer queries like "110 volt socket accepting grounding plugs"? Of course search engines can (and do) get such info, but afaik it doesn't require or use XSLT beyond extracting the plain text.

That is exactly the thought behind SGML/XML and its derivatives. XSL is kind of clumsy but very powerful and the most direct way to transform documents.

JSON-LD to me looks more like trying to glue different documents together, its not about the transformation itself.

> This is content-presentation separation at best.

The idea is the best, but arguably the implementation is lacking.

> Is JSON-LD just reinventation of this?

Yup. It's "RDF/XML but we don't like XML"

Reply on news.ycombinator.com