Linked Data and Semantic Processing Task Force

31 Mar 2017

See also: IRC log


Dave, Darko, Danh, Maria, Taki, Yingying, Achille


<scribe> scribe: Dave

<scribe> scribenick: dsr

Darko welcomes everyone to the task force meeting. We have two points in the agenda. 1) to review the charter and 2) to discuss use cases

Darko introduces Danh Le Phuoc (TUB). He is a member of the W3C group on spatial data on the web and has worked on the semantic sensor network ontology.

Danh: I am at the Technical University of Berlin and previously the DERI Insight centre

I’ve worked on applying semantic technologies for IoT for a number of EU projects

I am one of the editors for the W3C SSN ontology specification

There is lot of scepticism around RDF to dispel.

I joined the Web of Things IG early on, and expect to rejoin on behalf of TUB.

Dave: can you tell us more about people’s reactions to RDF?

Danh: a lot of EU projects have exploited Linked Data and ontologies, but there has been very little transfer into commercial products.

People tend to bypass RDF when working with constrained devices.

Darko: It may be very early for RDF in today’s products. Currently the focus is more on connectivity to the cloud and much less on interoperability where RDF and metadata should provide real value/. Let’s see!

Danh: schema.org is a success for the Web due to its ease of use, and we need to make it equally simple to adopt semantic technologies for the IoT.

Dave: I think that an emphasis on simplicity is really key to success

Darko: I agree too.

<DarkoAnicic> Pull request with proposed changes for the charter:

<DarkoAnicic> https://github.com/danicic/wot/blob/a0b9d22adde95d1aa8c9f51e53f7fb83ecd09c10/TF-LD_charter.md

<taki> Darko: I went to issues.

<taki> Darko: First part, introduction, I made no changes.

<taki> Darko: I added scope part.

<taki> Darko: Platform contract, I added second point, to align with ontologies. To ensure interoperability, at data model level.

<taki> Darko: Thing recipem extensibility of standardized vocabularies. Use cases and practices, how to enrich with semantics.

<taki> Darko: Constraints and validation. Integrity constraints and validating TD based on certain vocabularies and constraints.

<taki> Darko: Common activities is about interaction with other groups. I separated this from platform contract.

<taki> Darko: If we get another issue, we will update.

<taki> Darko: Let's discuss current issue.

<taki> DR: I found a typo that I can fix.

<taki> DR: Recipe. We need to clarify termonology.

<taki> Darko: We can use existing terms.

<taki> Darko: Nodered, IFTTT platforms use this term "recipe".

<taki> Darko: OCF, OneM2M can bring their module.

<taki> DR: Syntactic and semantics modularization.

<taki> DR: Need to consider concepts and how to name it.

<taki> DR: What kind of modularization technique have we used?

<taki> Danh: Vertical and horizontal. Saucer(?), schema.org avoid heavy schema.

<taki> Danh: Other overlays on top.

<taki> Darko: Modularization here, recipe is application template with respect to TD.

<DanhLePhuoc> +q

<taki> Darko: Quite simple template of TD. Standard vocabularies, we will use in that.

<taki> Achille: Syntactic and semantics. Recipe is not only about template but also orchestration of entities.

<taki> Achille: Semantics description of how to make them work together.

<taki> Achille: With right combination, output/input datatypes, accessor description.

<taki> Achille: It may be a bit different from modularization, but I like that concept.

<taki> Darko: Can you invite someone from OCF?

<taki> DR: We could ask.

<taki> DR: OCF and OneM2M.

<taki> DR: OCF airconditioner device. I reverse engineered their spec.

<taki> DR: Resource type, comments. Binary switch and temperature.

<taki> DR: There are syntactic combinations.

<taki> DR: Semantic receipe. How the switch is realized.

<taki> DR: I am happy to keep the name "recipe".

<taki> Darko: In TD, we already have this interface defined. properties, actions and events. We can annotate.

<taki> Darko: How these ingredients are combined together.

<taki> DR: We need to support both syntactic and semantic level modularization.

<taki> Danh: Semantic webservice.

<taki> Danh: It uses similar way.

<taki> Danh: composition to create new service.

<taki> Danh: Recipe here is different from there.

<taki> DR: I would like to enable application so it can discover devices at home.

<taki> DR: Based on semantic description.

<taki> DR: We need object model and semantic model.

<taki> Darko: We do not need to discuss solution at this point.

<taki> DR: We make change to "Semantic and syntactic templates".

<taki> DR: "recipe" is a top level name.

<taki> Darko: Do you want to do this?

<taki> DR: I will do, and we can refine it.

<taki> Darko: Next item.

<taki> Darko: Michael Koster raised it.

<taki> Darko: Everyone, please take a look at this issue.

<taki> Darko: Use cases, practices. Easily annotate descriptions. Common vocabularies. We need to show use case for this. Tools that are more user-friendly.

<taki> Darko: And validate against those vocabularies.

<MariaPoveda> which is the issue number?

<taki> DR: We need to this. Other several groups are looking at this issue. We need to collaborate.

<taki> DR: We need to demonstrate the benefits. Joint whitepaper is good.

<taki> MM: Automitive WG, Hydra community group. OpenAPI and Swagger. OCF is using swagger. Swagger does not support Linked Data now.

<MariaPoveda> Is this one https://github.com/w3c/wot/issues/312?

(the proposed W3C CG on smart homes and mobility, and the joint white paper on semantic interoperability organised by Paul Murdock)

<taki> Darko: We need to plan invited talks.

<taki> MM: Swagger group is most interesting to engage with.

<taki> Darko: OCF uses swagger?

<taki> MM: There is an emphasis on swagger lately.

Michael: also the W3C automotive WG and the folks working on swagger

<taki> MM: Instead of RAML, for various reasons.

<taki> Darko: Next point, semantic constraints and validation.

<taki> Darko: The idea is to show the benefits of semantic processing.

<taki> Darko: Semantic validation against certain model and constraints.

<taki> Darko: It can be quite helpful in many applications.

<taki> Darko: There are already works in different groups.

<taki> Darko: Existing techniques. Our proposal is quite helpful. Can show in Plugfests.

<taki> MM: Security. We are focusing on TD deliverable.

<taki> MM: Security reasoning.

<taki> DR: +1

<taki> MM: We also have security TF.

<DarkoAnicic> +1

<taki> Danh: Security and linked data. RDF and Linked data. Is it possible to provide better control?

<taki> MM: Interaction controls. Triples are better. RDF is useful for this.

<taki> MM: Composition of security propeties.

<DarkoAnicic> MM: Access control based on RDF.

<taki> MM: We should think about enabling this.

<taki> Darko: Can you raise this issue?

<taki> Darko: security control and access control using RDF.

<taki> ACTION: Michael MCool to raise an issue about security control and access control using RDF. [recorded in http://www.w3.org/2017/03/31-wot-minutes.html#action01]

<trackbot> 'Michael' is an ambiguous username. Please try a different identifier, such as family name or username (e.g., mcokus, mmccool, mvbekkum, mwesterm).

<taki> Maria: There is an issue.

<taki> Maria: Issue 306.

<taki> Maria: We need different focuses. What is the model. This is one. Entity relationship.

<taki> Maria: Another. Inference.

<taki> Maria: I prefer to have clear paths.

<taki> Darko: Thing description model is separated.

<taki> Darko: Validation based on Shapes. Inference based on Owl. We can capture them in one item.

<taki> Danh: I suggest to have best practice about reasoning, tranformation.

<taki> Danh: Investigation into readiness of technologies. LD tool chains.

<taki> Darko: Maria and Danh. Please raise those two issues.

<taki> Darko: We are already one minute late.

<taki> Darko: Dave's presentation, can we postpone to next time?

<taki> DR: OK.

<taki> Darko: You said you will make changes, and refine them. After that we will take a look at issues from MM, Maria and Danh.

<taki> MM: Timezone is wrong in calendar.

<taki> MM: Can we talk with Kaz?

<taki> Darko: Kaz updated wiki page.

<taki> Darko: It should be correct.

<taki> Darko: I hope this problem should not happen again.

<taki> MM: We need to fix it.

<taki> Darko: I will talk to Kaz.

<taki> DR: That would be helpful.

<taki> Darko: Let's talk in two weeks.

<DarkoAnicic> yes, I cannot create minutes

<DarkoAnicic> yes, I can send an email but I need the link for the minutes

<DarkoAnicic> ok, now it works - I will send the link

Summary of Action Items

[NEW] ACTION: Michael MCool to raise an issue about security control and access control using RDF. [recorded in http://www.w3.org/2017/03/31-wot-minutes.html#action01]

Summary of Resolutions

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.152 (CVS log)
$Date: 2017/03/31 15:14:42 $