- From: ☮ elf Pavlik ☮ <perpetual-tripper@wwelves.org>
- Date: Thu, 13 Nov 2014 00:24:47 +0100
- To: Erik Wilde <dret@berkeley.edu>, Owen Shepherd <owen.shepherd@e43.eu>
- CC: Markus Lanthaler <markus.lanthaler@gmx.net>, public-socialweb@w3.org
On 11/13/2014 12:12 AM, Erik Wilde wrote: > On 2014-11-13, 00:04, ☮ elf Pavlik ☮ wrote: >> On 11/12/2014 11:50 PM, Owen Shepherd wrote: >>> Whether we define our own types as aliases of others - e.g. "as:Person >>> owl:sameType vcard:Individual" - or just refer to others is kind of >>> immaterial, though I'm in favor of the aliasing because that >>> implementers don't have to think "People come from VCard, Events come >>> from iCalendar, businesses come from Org..." >> I see limitations with depending to much on inference, for example >> unhosted[1] apps which run fully in a browser as for today don't have >> any reasoners available (javascrip) :( > > it would be very helpful to make a very clear (and hopefully > subsequently stable) decision on whether AS implementations are expected > to operate on the RDF data model, and specifically use inference, or > not. if that's the case, we will explicitly exclude the vast majority of > implementations and implementers out there. I would say that we can base it on RDF data model so *those who chose to* can take full advantage of it. Still if someone *chooses to* ignore @context, then this implementation will treat the data as plain old JSON, which uses unmapped strings. Constructs like "@type": ["Person", "foaf:Person"] and similar don't force anyone to treat them as RDF. I think we should still discourage (maybe even forbid) using JSON features not supported by JSON-LD. So far I can only think of array of arrays (GoeJSON-LD case).
Received on Wednesday, 12 November 2014 23:26:56 UTC