Re: Tomorrow the 7th TREE CG call!

Hi all,

Thanks Xueying, Thomas, Wout and Julián for attending the call today!

  *   The slides: https://docs.google.com/presentation/d/1CqpwV9ScP9L44QfBYASQZcTGDMMra6U8yUa0MzU70gU/edit#slide=id.g28b1beea129_0_58
  *   The decisions:
     *   We’ll start introducing semantic version and a changelog from now on, so that implementations can describe which version of the spec they used. It must be clear however that our group aims to produce a 1.0 specification that already now won’t introduce any breaking changes with existing implementations out there.
     *   The next roadmap:
     *   Prioritizing issue #85 now<https://github.com/TREEcg/specification/issues/85> - today we’ve introduced the background (see slides) and what extra feature would be needed. There were still different opinions about what would be needed and we’ll continue the discussion on github. We’ll pick up from where we left off on the next call.
        *   Sander wants to introduce a specific kind of relation (subsetrelation) that if you would only follow these links, that you can ignore the other relations if you only want to go deeper in the search tree. If I understand correctly, he sees use cases where, when you found a very specific tree:Node that adheres to all constraints to get there, that you’d want to server to give a link to a neighbour that adheres to a different set of constraints.
           *   Pieter’s 2 cts here was that he didn’t see any usefulness in that, as the client will already have discovered the links to these neighbours in an earlier node, and that if the client is interested in something next to this, it should have simply indicated that from the beginning. I thus don’t see the need for a subsetrelation as I believe all relations should be subsetrelations as long as a client prunes the ones it already saw.
        *   Pieter instead wants to introduce a new property as an alternative to void:subset to indicate a non-root-node that will not link out of its own subgraph, and won’t get inbound links into its subgraph.
        *   Thomas indicated that he certainly needed backlinks to stay legal in the spec as-is, which nobody contested and that I assume consensus on.
     *   Next steps on the roadmap once issue 85 is resolved:
        *   We need to raise the developer experience of the spec -- we should plan some proofreading sessions to make it more developer friendly than competing specs
        *   Talk about removing unnecessary features and clutter from the spec, such as the conditional imports system that was largely deprecated thanks to a better member extraction algorithm
        *   We need to start talking about our standardized test suite for conformance to TREE
     *   The Eighth TREE CG call will be held on the 25th of October at 10 am CET - put this in your calendar now! Same link!

Kind regards,

Pieter

________________________________
From: Pieter Colpaert (UGent-imec)
Sent: Tuesday, October 10, 2023 5:13 PM
To: public-treecg@w3.org <public-treecg@w3.org>
Subject: Tomorrow the 7th TREE CG call!

Hi all,

Tomorrow we’re going to have the 7th TREE CG call at 15:00 CET on Teams<https://teams.microsoft.com/l/meetup-join/19%3Ameeting_NDkwMWZlMzMtM2FjZi00MjZhLTlhZTMtNjAwMjU5Yjc3YWVi%40thread.v2/0?context=%7B%22Tid%22%3A%22a72d5a72-25ee-40f0-9bd1-067cb5b770d4%22%2C%22Oid%22%3A%22074b6191-940e-49de-964e-f2919f3f8501%22%2C%22MessageId%22%3A%220%22%7D>

Last time we said we’d proof-read the spec (always a good idea), do issue prioritization and talk about plans towards a test suite. I’d actually immediately would like to prioritize one specific issue to be discussed tomorrow, and I’ll prepare some slides to sketch the background on TREE traversal and the design of the relations. The issue and proposal was created by Xueying Deng and Sander Van Dooren: #85<https://github.com/TREEcg/specification/issues/85>

Looking forward talking to everyone tomorrow again!

Kind regards,

Pieter