- From: Tomasz Pluskiewicz <tomasz@t-code.pl>
- Date: Fri, 25 Sep 2015 21:23:47 +0200
- To: public-hydra@w3.org
On 2015-09-25 15:03, Asbjørn Ulsberg wrote: > 2015-09-25 14:33 GMT+02:00 Ruben Verborgh <ruben.verborgh@ugent.be > <mailto:ruben.verborgh@ugent.be>>: > > That's fine. I don't want to break that compatibility. > > I.e., if Alice has friends Bob, Charly, and Dean, then this is > :alice :hasFriend :Bob. > :alice :hasFriend : Charly. > :alice :hasFriend : Dean. > > > This is also fine. But this is on an abstract, RDF level. I want it to > also make sense in the bits and bytes that produce the sensible JSON > most people will understand. The RDF part of JSON-LD is not something > that most people having (or wanting) to deal with it will even know the > first thing about what is or how works. It's mostly just very convenient > that we finally have a format for expressing relations that doesn't need > a scientist to understand. > I don't agree. JSON-LD naturally stems from the JavaScript land, where JSON also originated. JavaScript is often used in a way where a given parameter accepts either a single element or an array. In such cases I would expect a singular form. For example in knockout validation it is possible to supply a single custom validator or an array [1]. . Plural would mean that an array is always expected, which is not true. To back it up, consider a JSON-LD document, where a property is defined as "@container": "@set" (or list). Both a single element or JSON array produce equal expanded JSON-LD and triples. See playground examples [2] and [3]. Thus I think that it isn't necessary to do what you propose and may arguably be less confusing than you fear. Just my two cents. Regards, Tom [1]: https://github.com/Knockout-Contrib/Knockout-Validation [2]: http://tinyurl.com/ndhoq5t [3]: http://tinyurl.com/nj2b2rl
Received on Friday, 25 September 2015 19:24:39 UTC