Re: Setting milestones (Re: marathon)


On 04/04/2022 23:26, Gregg Kellogg wrote:
> This use case might be better served by the Linked Data Patch 
> Format [1] (wink wink PA),
ah! Nice to see that somebody remembers that ;)
> which is more specifically intended for this, and doesn’t require the 
> invention of new semantics for something like solid:where/insert/delete.

I guess it all depends whether you have an N3 parser available, as 
opposed to building a specific parser for LDPatch...

Note that a lot of work on LDPatch went into the handling of lists 
(which was our main selling point compared to using SPARQL as a patch 
language).

In order to provide an equivalent ease of use for lists, N3 patch must 
support the builtins from the list: namespace [1]. More generally, I am 
curious to know which builtins, if any, are required by the N3 patch 
protocol.

[1] https://w3c.github.io/N3/ns/#vocab_list_append


>
> Gregg Kellogg
> gregg@greggkellogg.net
>
> [1] https://www.w3.org/TR/ldpatch/

>
>> On Apr 4, 2022, at 1:53 PM, Jos De Roo <josderoo@gmail.com> wrote:
>>
>> There is a crucial piece of N3 for the specification of the Solid 
>> Protocol
>> more specifically at 
>> https://solidproject.org/TR/protocol#writing-resources

>>
>> It uses cited formulae with quickvars (?x), so it is beyond N3-Lite
>> but it actually is an excellent and practical use case.
>>
>> That said, EYE suffers to roundtrip (with --pass) the example in
>> https://solidproject.org/TR/protocol#writing-resources

>> The n3p code at
>> https://github.com/josd/n3p/blob/33dde4db84c48467cd9e188744e8f34672675c7e/examples/patch.n3p

>> roundtrips fine at
>> https://github.com/josd/n3p/blob/33dde4db84c48467cd9e188744e8f34672675c7e/result.n3p#L103

>>
>> jos
>>
>> -- https://josd.github.io

>> <https://josd.github.io/>
>>
>>
>> On Mon, Apr 4, 2022 at 12:43 PM Pierre-Antoine Champin 
>> <pierre-antoine@w3.org> wrote:
>>
>>
>>     On 04/04/2022 09:40, Miel Vander Sande wrote:
>>>     Hi Jos, all
>>>
>>>     Thanks for running this marathon; I think we all very much
>>>     appreciate EYE and the updates that it got because of this.
>>     +42 :-)
>>>     I'm clearly one of the 33 dormant members of this group, mainly
>>>     because I'm really unqualified to partake in the discussions (I
>>>     see N3 as a tool, unwary of the insides), but may I ask:
>>>     after 42 months, is there a group report in sight? Is there a
>>>     milestone planned?
>>
>>     That's a fair question, and we acknowledged during the call last
>>     week that, unfortunately, we are not there yet. Defining a clean
>>     definition of N3's semantics is getting in the way, and
>>     unfortunately eating a lot of our bandwidth. I probably have my
>>     share of responsibility in that latter point, and I apologize for it.
>>
>>     Here is an idea that is maybe silly, but maybe can help us make
>>     progress : we could focus, for a while, on a strict subset of the
>>     N3 language (coined e.g. N3-Lite), and try to get a
>>     self-sufficient CG report on that subset. Then we could try to
>>     grow this subset (possibly in several incremental steps) until we
>>     cover N3 entirely.
>>
>>     The subset I have in mind is the following (but that's open for
>>     discussion):
>>
>>     - quoted graphs are disallowed, expect as the head or body of rules
>>     - rules are not encoded as triples, but handled at their own level
>>     - quickvars (?x) are only allowed in rules
>>     - no explicit quantification
>>
>>     I think that defining the semantics of this subset should be
>>     relatively easy (compared to full N3), and that it could be done
>>     in such a way that all existing N3 implementation already comply
>>     with the semantics.  We would therefore have a first level of
>>     interoperabilty formally specified.
>>
>>     Focusing on N3-lite, we could also come back to our work on
>>     builtins (although the most complex ones, such as log:semantics,
>>     log:includes, or log:forAll would not be part of N3-lite).
>>
>>     I know that this is reminiscent of the different profiles of OWL,
>>     which some people in this group don't quite like. Maybe to avoid
>>     this issue could we decide that N3-Lite needs to disappear once
>>     N3 "full" is properly specified. In any case, I think that
>>     N3-Lite could be a useful stepping stone.
>>
>>>     And is there anything non-expert community members can do to help?
>>
>>     Hopefully, the discussions on N3-Lite will be more accessible,
>>     and allow for a wider group to engage. At least, that's one of
>>     the goal of this proposal.
>>
>>       pa
>>
>>>
>>>     Best,
>>>
>>>     Miel
>>>
>>>     Op ma 4 apr. 2022 om 02:01 schreef Jos De Roo <josderoo@gmail.com>:
>>>
>>>         Hi all,
>>>
>>>         This group started about 42 months ago
>>>         https://lists.w3.org/Archives/Public/public-n3-dev/2018Nov/

>>>         and for me it really feels like a marathon coming to an end.
>>>         During this past 42 months we had lengthy discussions
>>>         https://docs.google.com/document/d/1A3HAUhjaVnnJ6yVbFAvIBRJQjUY9aFlQ2_bGxkD0mnE/edit

>>>         and as a result for the eye reasoner there were 244 releases
>>>         https://github.com/josd/eye/blob/master/RELEASE

>>>
>>>         For me what really remains is n3p which is the eye
>>>         intermediate p-code and from now on
>>>         I will focus on https://github.com/josd/n3p as a
>>>         https://knows.idlab.ugent.be/team/ member
>>>         and stay quiet in this N3 community group like most of the
>>>         other 33 members.
>>>
>>>         Thanks and kind regards,
>>>         Jos
>>>
>>>         -- https://josd.github.io

>>>
>

Received on Tuesday, 5 April 2022 19:14:53 UTC