W3C

– DRAFT –
WoT and NGSI-LD

17 February 2025

Attendees

Present
Franck_Le_Gall, Juanjo_Hierro, Kaz_Ashimura, Ken_Zangelin, Martin_Bauer, Michael_McCool, Zane_Durkin
Regrets
-
Chair
McCool
Scribe
kaz

Meeting minutes

Minutes

Feb-3

McCool: (goes through the minutes)

typos fixed; approved

Cancellations

McCool: cancellation on Feb 24
… the next meeting will be March 3
… we can keep the meeting weekly but can cancel the meeting if people are not available

Activities

Resources

resources.md

McCool: I've been reading the NGSI-LD spec
… basically, NGSI-LD discover the database
… what is interesting is that WoT Directory also finds the TD as the database
… so very similar design with each other

Martin: we have open APIs

<martin> NGSI-LD OpenAPI is available here: https://forge.etsi.org/rep/cim/ngsi-ld-openapi

McCool: ok
… the difference from my viewpoint is
… the WoT Discovery is very simple
… we can get Thing Description, which is JSON-LD
… what would be the technical requirements?
… 1. context management
… we can use nested contents

McCool: wondering about the compatibility
… 2. TD using special Things
… one of those is default values
… normalize the TD to formalize the data

Ken: we also have default values too

Martin: but maybe a bit different level...
… we use JSON-LD as representation
… what is our structure looks like

McCool: default value is just one example
… we should investigate the differences

McCool: a set of nodes to be included

<martin> JSON-LD is a serialization of RDF (as abstract model) - RDF/XML is another serialization of RDF

Juanjo: definitely interesting
… digital twin characteristics
… we could generate TD based o the TM template
… have to discuss the digital twin class

McCool: TD describes entities
… could be a digital twin
… what TDs bring to the table include data modeling languages
… describing the payload
… also describe affordance, i.e., property, action and event

Kaz: tx for your investigation, McCool
… we should now start to describe what is done one the NGSI-LD side and what on the WoT side
… based on some concrete use case
… to clarify what the NGSI-LD broker and the WoT Directory do there

McCool: yeah
… we should capture the use cases based on our understandings

Issue 8 - Capture Use Cases

<jhierro7> This presentation brought the idea I had commented:

<jhierro7> https://docs.google.com/presentation/d/1-RI8q1hhnRqaZDwvxnpKnaaxQ0kMGtME/edit?usp=sharing&ouid=115453639179713725602&rtpof=true&sd=true

McCool: should add a link for that also to the resources.md

Juanjo: also a PR to add that to the resource.md

PR 9 - Adding presentation in section about Presentations

McCool: (merged)
… (shows slide 7. Compatibility of NGSI-LD with Thing Models/Descriptions)
… (also slide 12. Compatibility of NGSI-LD with Thing Descriptions)

Juanjo: another PR about the idea of TM as template

PR 10 - Update section on Thing Models as Templates

McCool: (merged)
… let's think about the next steps now

example/README.md

McCool: (shows examples/README.md)
… would create a new Issue to document technical gaps
… (adds questions)
… Does "type" abstraction in NGSI-LD map onto the ability of TMs to describe classes of potential TD instances?
… TMs used for describing the properties of entity types?
… conceptually, TMs describe classes
… note that TM doesn't have network interface

McCool: What is the subclass structure?
… TMs for everything or for subclasses of "entities with affordances/network interfaces"?

Juanjo: TM describes what to be exposed

McCool: technically, TM is a class of template
… but it's a templating mechanism
… (revisits the 3 questions issued within the GitHub Issue)
… (and adds another question)

<jhierro7> In my view, Things Model templates allow to describe the characteristics of any instance of an entity type (Things class if you want)

<jhierro7> Things Description can be derived from those templates, by adding forms how those characteristics can be accessed using specific protocols.

McCool: Should TMs/TDs we attached as annotations or be included in the NGSI-LD type structure?

<jhierro7> In such case, TM can be used to describe characteristics of Things = Entity Types

<jhierro7> and then, we may specify how TDs can be derived where NGSI-LD is added as form to access characteristics

McCool: TMs may so be more "dynamic" than (what seem to be) the pre-defined "types" used by NGSI-LD
… would like to understand what is possible at the moment

Kaz: some discussion about how to generate TD instance from TM last week during the TD call
… this discussion about TM and NGSI-LD should be useful for that (=TM vs TD) discussion also
… so would suggest we clarify our expectation for TM a bit more

McCool: yeah, we should have more discussion
… let's add the link about the discussion from the TD TF too

Other issues and PRs

Remaining Issues

Remaining PRs

Juanjo: would like to work on some more exercise

McCool: let's take a few TMs/TDs and NGSI-LDs to see the differences

<jhierro7> The exercise of taking an entity type description using the WoT Models Template, and show how Thing Descriptions can be derived using NGSI-LD as form.

<jhierro7> (form for access)

McCool: Review and comment on new issues
… Test importing TDs into NGSI-LD broker
… Generate a concrete example of using TM/TDs for Digital Twin modelling

AOB

McCool: next meeting on March 3
… will update the agenda
… we'll keep the meeting as weekly
… but will send a cancellation notifications based on the actual schedule

[adjourned]

Minutes manually created (not a transcript), formatted by scribe.perl version 242 (Fri Dec 20 18:32:17 2024 UTC).