Re: Observations on WebID definition and specification

+1

On Mon, Feb 5, 2024 at 4:43 PM Nathan Rixham <nathan@webr3.org> wrote:
>
> > I would like all members of the group to be able to agree this. To temporarily strip the MUST mediatype, and defer all discussion of it until everything else is complete.
>
> What's the best way to propose this defer approach as something actionable - as we're already cycling the same age old conversations.
>
> On Mon, 5 Feb 2024, 13:39 Martynas Jusevičius, <martynas@atomgraph.com> wrote:
>>
>> Jacopo,
>>
>> I don't completely get the ecosystem argument.
>>
>> Mandating serializations in a spec does not automagically make parsers
>> for those serializations appear in the RDF ecosystems, does it?
>>
>> Content negotiation is a transparent HTTP mechanism for finding an
>> intersection between server-supported and client-supported
>> serializations. Overriding it by mandating serializations explicitly
>> only breaks this transparency, and does nothing to increase
>> interoperability IMO.
>>
>> Martynas
>>
>> On Mon, Feb 5, 2024 at 4:29 PM Jacopo Scazzosi <jacopo@scazzosi.com> wrote:
>> >
>> > Hey Martynas,
>> >
>> > Answering personally, I don’t think that there’s anything justifying a different media type treatment. I do think, however, that the RDF ecosystem as a whole poses a significant challenge to implementors due to the amount of formats that are in use within it. I would, therefore, make the same considerations WRT any other foundational (or would-be foundational) RDF spec. I don’t do so because I only participate in this CG :).
>> >
>> > Speaking more practically, the state of RDF parsers and serializers is nowhere near as developed once one looks outside of the Java and NPM ecosystems. Even whithin these two, and I’m personally more familiar with the NPM one, supporting RDFa, Turtle and JSON-LD often pushes dependencies past the limits imposed by corporate processes (license clearing, to make an example). Furthermore, formats like full-fledged JSON-LD and RDFa are rather unfriendly to low-power applications.
>> >
>> > So yeah - no particular difference that I can see. However, and I’m growing more confident every time I think about it, I do think that removing media type requirements could well lead to format convergence, if perhaps  counterintuitively. I’d be willing to take that risk.
>> >
>> > Best,
>> > J.
>> >
>>

Received on Monday, 5 February 2024 13:54:05 UTC