Re: [EXTERNAL] Re: Straw-poll: Issue 410 / Allowing list syntax

I prefer 1 with the restriction others have noted: Only the primitive NodeKinds should be in the lists.

I see some performance argument to allowing lists to be done in different orders than we’d want to spell out in IRIs: Users might have a tuned order they know would pass quickly the majority of the time, based on knowledge of their data behaviors.

--Alex


From: Yousouf Taghzouti <yousouf.taghzouti@inria.fr>
Date: Friday, July 18, 2025 at 2:47 AM
To: David Habgood <david@kurrawong.ai>
Cc: Robert David <9427084@gmail.com>, Holger Knublauch <holger@topquadrant.com>, Public Shacl W3C <public-shacl@w3.org>
Subject: [EXTERNAL] Re: Straw-poll: Issue 410 / Allowing list syntax
I also prefer option 1, but with the proposed restriction to the four base/single kinds.

Yousouf

________________________________
De: "David Habgood" <david@kurrawong.ai>
À: "Robert David" <9427084@gmail.com>
Cc: "Holger Knublauch" <holger@topquadrant.com>, "Public Shacl W3C" <public-shacl@w3.org>
Envoyé: Lundi 14 Juillet 2025 13:38:09
Objet: Re: Straw-poll: Issue 410 / Allowing list syntax
I prefer option 1 with the proposed restriction to the 4 base/single kinds.

Likewise. Although lists can be painful to work with I do find it cleaner and already need to handle them elsewhere.

David Habgood

Knowledge Graph Senior Developer

KurrawongAI



[Image removed by sender. emailAddress]
david@kurrawong.ai<mailto:david@kurrawong.ai>
[Image removed by sender. website]
https://kurrawong.ai<https://kurrawong.ai/>



[Image removed by sender.]


On Monday, 14 July 2025 at 19:29, Robert David <9427084@gmail.com> wrote:
I prefer option 1 with the proposed restriction to the 4 base/single kinds.

Am Sa., 12. Juli 2025 um 10:49 Uhr schrieb Holger Knublauch <holger@topquadrant.com<mailto:holger@topquadrant.com>>:


> On 11 Jul 2025, at 5:54 pm, Eliana Papoutsoglou <Eliana@y.digital> wrote:
>
> Hi all,
>
> As discussed in our last meeting, this is a straw poll to help us assess the group's position on PR #410: "feat: #407 add support for sh:TripleTerm to sh:nodeKind and allow lists."
> https://github.com/w3c/data-shapes/pull/410

>
> The main point of contention is whether lists should be allowed.
>
> On one hand, lists offer a more user-friendly syntax that can better accommodate new use cases compared to the current sh:or. On the other hand, allowing lists may introduce performance issues.
> If the group does decide in favour of allowing lists, there will be additional options to explore (such as simplifying use through node expression functions). However, for this vote, I would like to keep the focus on whether they should be allowed in the spec at all.
>
> This is a straw poll to gauge preferences. Please indicate your preferred option:
>
> 1) Lists should be allowed.
>
> 2) Lists should not be allowed.

I prefer 2 due to the added complexity (cost for tools/learning), overlap between syntax options and the lack of convincing use cases for ORs including TripleTerms.

But I could live with option 1 if it's limited to the 4 base kinds and the majority of people prefers that.

Holger


>
>
> The poll will remain open for 10 days, until the WG meeting on July 21st at 14:00 CET. Please submit your response before then.
>
>
> Kind regards,
> Eliana
>
>