Re: Reified triple syntax

maybe we can postpone this discussion about syntax a bit?
andy and gregg pulled a little stunt and whipped up some changes to the syntax without much involvement of anybody else, and suddenly it was the new syntax. that was a bit surprising, and probably also a bit confusing, but i don't care too much about it at this point. 

that said:

Am 20. September 2024 09:36:30 MESZ schrieb Franconi Enrico <franconi@inf.unibz.it>:
>Syntax is important since it is the interface of the language with users.

i guess we all agree on that

>Choices should not be made favouring ease of parsing vs ease of understanding.

choices should be made favoring ease of understanding, but there is of course also a limit to what parsers can be expected to digest

>Having the reifier after the triple is the exact opposite of what most of the people felt about not having triple terms in subject position!

that to me seems really very unrelated. 

>Make up your mind and decide whether reifiers should syntactically appear before or after the triple term, and be consistent between turtle and ntriples.

i disagree. turtle-star is the user interface, ntriples is the programmers interface. 

.t

>—e.
>
>> On 20 Sep 2024, at 00:50, Andy Seaborne <andy@apache.org> wrote:
>> 
>> Forked thread.
>> https://lists.w3.org/Archives/Public/public-rdf-star-wg/2024Sep/0062.html
>> 
>>> On 19/09/2024 11:17, Franconi Enrico wrote:
>>>> On 19 Sep 2024, at 17:44, Andy Seaborne <andy@apache.org> wrote:
>>>> The syntax for optionally naming the refiier in an occurrence changed.
>>>> 
>>>> Occurrence syntax: << :id | :s :p :o >> :pp :oo
>>>> ==>
>>>> Occurrence syntax: << :s :p :o ~ :id >> :pp :oo
>>> 
>>> Let me voice my personal opinion about this change: I don’t like it,
>> since it it much less legible to me.
>>> If you really insist on having the “~” symbol, I’d rather prefer:
>>>  << :id  ~ :s :p :o >> :pp :oo.
>>> since it emphasise in a direct way that the denotation of that term
>> is “:id”.
>>> —e.
>> 
>> Position:
>> 
>> Having the reifier id at the end is the same style as annotation.
>> 
>> # Reified triple declaration
>> << :s :p :o ~ :r >> .
>> 
>> :s :p :o ~ :r .
>> :s :p :o ~ :r  {| :q :z |} .
>> 
>> Having it "pre" in one case and "post" in the other is a bit strange IMO.
>> 
>> On a technical level, it keeps the grammar requirements simple.
>> 
>> Seeing "<< :r", the parser can't tell if that is a subject or a reifier id. Initial placement is possible at the cost of more complicated rules or a lookahead of 2+ which limits the implementation tooling available.
>> 
>> Symbol:
>> 
>> '|' is visually confusing for SPARQL.
>> 
>>    Andy
>>

Received on Friday, 20 September 2024 08:34:35 UTC