W3C home > Mailing lists > Public > public-rdf-star@w3.org > January 2021

Re: Summary: Annotation Syntax Proposals

From: Andy Seaborne <andy@apache.org>
Date: Fri, 22 Jan 2021 08:49:50 +0000
To: public-rdf-star@w3.org
Message-ID: <f964404f-dff1-47bb-32a1-7d188fe9976b@apache.org>


On 21/01/2021 16:44, Pierre-Antoine Champin wrote:
> 
> On 21/01/2021 17:17, Miel Vander Sande wrote:
>> No real objection here. I can agree with the points made.
>> That said, I'd still not use { }.
> The main argument in favor of curly brackets, as I understand, is that 
> they are not used in Turtle, so they make annotations easier to 
> differentiate from other syntactic constructs. That is, in Turtle*...
> 
> I agree with you that, in the bigger picture (i.e. SPARQL*, TriG*), this 
> creates an irregularity. I would also rather use square brackets for 
> this. May be [| ... |]  ?
> 
>> If any, just use :s :p1 :o1 | :source :URL1 |.
> 
> That would make embedded annotations tricky to parse (automatically, not 
> to mention visually).
> 
> :s :p1 :o2 | :a :b | :c :d |, :e | :f :g | |.
> 
> :-/

Right - having same start and finish does not work. The second "|" can 
be either a start or a finish.

Even if we say that nesting is not expected and need not have much 
design weight, designing it out makes the impact on the grammar wider.

 From the comments about {} being triples, then trying to make [..] work:

[| ... |]

or |[   ]|
or @[   ]

(note langtags have at least one alphabetic character so @[ is not a 
clash with an existing langtag)

are possibilities.

     Andy
Received on Friday, 22 January 2021 08:50:05 UTC

This archive was generated by hypermail 2.4.0 : Friday, 22 January 2021 08:50:07 UTC