Re: webid serializations consensus 2023

Hi all,

some comments:

* personally, I would like to see a client specification, i.e. 
specifying what a WebID reader MUST accept and be able to parse. I see 
it more helpful for adoption, if the publishers are unburdened and MAY 
choose the format they like the most. This would also improve 
validation, i.e. a standardized client can give a definite yes or no.  I 
did some work with https://www.unicode.org/uli/ in the past and I had 
the feeling that Unicode and ULI were always emphasizing that parsing is 
key, not producing.

* supplemental to this, I can report that turtle syntax causes some 
problems in the wild . We asked the DBpedia Community to make WebIds on 
github.io using webid.ttl (e.g. kurzum.github.io/webid.ttl) and there 
actually was a lot of heterogeneity. One problem, for example, was 
my:prop [ <a> "sth" ] , because some parsers require .  and some not 
like `my:prop [ <a> "sth" .  ] .`   Sorry, I can't remember specifics. 
Providers were testing with one relaxed parser and later retrieval & 
parsing failed with another (more strict). I would argue, however, that 
requiring (MUST) a provider to support more formats can lead to an 
increase in potential errors. Worst case none are correct due to focus 
loss or complexity of content negotiation rules. Incidentally, I spent 2 
hours earlier today debugging  Nginx redirect rules, which is not 
trivial and only succeeded with the help of ChatGPT.

* JSON-LD can be made to look like plain JSON with the right context. 
This ideally means 1. no datatypes, 2. no `vocab:`, 3. no full URIs in 
keys. See example below. I consider this a great plus for schema.org, 
although some providers put "@context" : "https://schema.org" instead of 
"https://schema.org/" , which I read as "we never tested this with a 
json-ld parser". Otherwise, there is the flattened form with "@graph" .

-- Sebastian

{
    "@context" : "https://downloads.dbpedia.org/databus/context.jsonld",
    "@id" : 
"https://dev.databus.dbpedia.org/denis/onto_dep_projectx/dbpedia-ontology/2023-03-14",
    "@type" : "Version" ,
    "description" : "Registered a version of the DBpedia Ontology into 
my account. \n\nUsing markdown:\n1. This is the version used in [project 
x](http://example.org) as a stable snapshot dependency \n2. License was 
checked -> CC-BY",
    "distribution" : [
       {
          "@type" : "Part",
          "dcv:sorted" : "true",
          "dcv:type" : "parsed",
          "downloadURL" : 
"https://akswnc7.informatik.uni-leipzig.de/dstreitmatter/archivo/dbpedia.org/ontology--DEV/2021.07.09-070001/ontology--DEV_type=parsed_sorted.nt"
       }
    ],
    "license" : "http://creativecommons.org/licenses/by/4.0/",
    "title" : "DBpedia Ontology"
}

On 6/15/23 17:34, Melvin Carvalho wrote:
> https://github.com/solid/solid-wg-charter/issues/39
>
> The consensus for WebID serializations ~10 years ago was:
>
> A WebID Profile Document is a Web resource that MUST be available as 
> text/turtle [turtle], but MAY be available in other RDF serialization 
> formats (e.g. [RDFA-CORE]) if requested through content negotiation.
>
> Since then there have been about a decade of discussions. We should 
> probably make the views of the CG known before it is handed off to the WG.
>
> JSON-LD has matured and gained significant adoption in the intervening 
> time period, and IMHO, should be a primary serialization.
>
> A starting point IMHO reflecting consensus could be:
>
> A WebID Profile Document is a Web resource that MUST be available as 
> text/turtle [turtle], MUST be available as application/l+json 
> [json-ld], and MAY available in other RDF serialization formats if 
> requested through content negotiation
>
> ie
> - Turtle MUST (+1)
> - JSON-LD MUST (+1)
> - others MAY
>
> I am unsure if we should vote and tally them, or preferably avoid 
> voting all together and agree unanimously.  If anyone disagrees, they 
> should be heard tho.
>
> Minor details:
> - JSON-LD can also be delivered in a structured data island
> - I removed the reference to RDFa, as I Think it's superciced, some 
> may be agaisnt that, but I think we can establish group consensus there
>
> I think handing that off to the WG would be a win all round.  Worth 
> adding your view, as it might be the last time we get a say on WebID 
> in this group.

Received on Tuesday, 20 June 2023 22:25:48 UTC