Merge pull request #131 from w3c/ege-meetup11

Meetup 11 Resources

Minutes/2023/2023_12_12-minutes.md

@@ -0,0 +1,207 @@
+Meeting Minutes
+===
+
+### info
+
+:date: **Date:** 12 December 2023
+
+### :bust_in_silhouette: Participants
+
+- Ege Korkan
+- Cristiano Aguzzi
+- Mikkel Haggren Brynildsen
+- Martin Fuhry
+- Luca Barbato
+- Christian Winkler
+- Minyong Li
+- Felix Paulini
+- Tatiana Mangels
+- York Schmidtke
+- Øivind
+- Ali Daher
+- Ganesh Ramanathan
+- Daniel Peintner
+- Thomas Hubauer
+- Bernd Utz
+- Mahda Noura
+- Michael Lagally
+- Josh Cohen
+- Luca Roffia
+- Michael Koster
+
+### :scroll: Agenda
+
+- News from the CG
+- Introduction of the Speaker
+- Presentation
+- Q&A
+- Discussion on Ontologies
+
+:writing_hand: **Scribe:** Cris
+
+:computer: **Slides:** https://docs.google.com/presentation/d/1r_7PrstswVoU_1awFvrcfa6tuQRzm76coDZkPfuY68k/edit?usp=sharing
+
+### Introduction
+
+Ege: Welcome everyone to the 11 Online Meetup. you already know the agenda.
+
+Ege: we have a bunch of links on the slides that you can use to keep in touch with us
+
+Ege: we have some progress to the collaboration with the WG group. Please use Github to follow the discussion.
+
+Ege: We are also moving on with our Tutorial website for onboarding on the WoT ecosystem.
+
+Ege: If you are interested in the home automation we have a focus group dedicated to that.
+
+Ege: we are also planning a next meetup for the next year about robot control.
+
+Ege: if you want to stay in the loop feel free to join our group. It is free and there are no strings attached.
+
+### Speaker Introduction
+
+Ege: Today we have Mikkel at Grundfos. He is Chief Data Scientist and specialized in Data & AI. Thank you for joining us
+
+Mikkel: thank you.
+
+### Presentation
+
+Mikkel: Hello, I work as data scientist and ontology. I'm also involved in standardization activities. Today I'm reporting the latest activities on our efforts in the standardization work for IoT.
+
+Mikkel: there are already a lot of contributors.
+
+Mikkel: we created an Industrial Data Ontology. It is an ISO and it is for represent industrial data and assets. It is based on other standards and it has been used in different industrial projects.
+
+Mikkel: used a lot in oil gas but also in discrete manufacturing. IT has 54 classes and 88 object properties. It is not huge.
+
+Mikkel: use case, you see a pilot in a ranch in Arizona. We have solar panels that provide power to a water pump. There is a need of intelligence about when to pump. We build a cloud solution that is able to understand when to pump.
+
+Mikkel: Hardware and Software team are far apart. We need a common understanding of the two world to build a great cloud solution.
+
+Mikkel: What solutions are there to move data from the sensor to cloud?
+
+Mikkel: WoT was useful, we didn't implement it fully. But we can construct a WoT TD from our database if needed. It used to tell people how to use the device without going into detail about how it is implemented in the device.
+
+Mikkel: in this use case the IDO model we wanted to meet engineers needs. We wanted to do this at scale. We have hundreds of different types of pumps. There are too many teams that need to cooperate when building a fully cloud solution from ground up.
+
+Mikkel: how can we value the knowledge of the developers able to connect hardware to the cloud? We wanted to answer this in this IDO use case.
+
+Mikkel: We started from the liquid temperature (it is a software estimation). Basically the software has to estimate the liquid temperature using the list temperature.
+
+Mikkel: Once estimated you have to serve the temperature over the network. Note that this is an official use case submitted to the ISO community.
+
+Mikkel: It is important to know the resolution of the data.
+The data resolution here is how to represent a continuos number in software.
+Considering using 4 bits to store temperature. It makes sense because you can make the pump cheaper. You should not waste memory.
+
+Mikkel: The calibration of the sensor is also important to keep the memory usage low.
+
+Mikkel: if you want to model that in a TD you get what is shown in slide 9. Multiple of is important.
+
+Mikkel: the step is important to correctly make users react to real changes. Any smaller changes are just noise.
+
+Mikkel: what you see is the full model, we have some hardware or physical artifacts.
+We have some software and activities modelled too. We also the data and how it move through the lifecycle. We have metadata on data points that tells has the resolution of the data. We store data as sub samples with larger step sizes in the devices. Last activity is the one where we estimate the temperature and the step size.
+There is where we use WoT multipleOf. The general rule is that you should not see the step sizes should not get smaller but only bigger. We can put the all graph in a RDF storage and validate it using SPARQL queries.
+
+Mikkel: this use case demonstrate how you can use IDO in WoT TDs. We can now communicate better between the teams. when everything is in place we can now deploy hot fixes and not just PDF documents. We can even extended it.
+
+Mikkel: that's all you can find my contacts in the slide.
+
+### Q&A
+
+Ege: we have been wondering if anybody out there is using our ontology. I think you demonstrate that you are using the ontology. Is that true?
+
+Mikkel: yeah I think we showed how you can validate one part of the file. We also tag the properties with IDO tags.
+
+Ege: this is perfect example on how we encourage to do. In the property graph you showed are importing the TD ontology?
+
+Mikkel: we use SSN and WOT ontology. We use WoT ontology directly. 
+
+Ganesh: Dealing with electrical mechanical system. We faced the issue when if a device expose its own Thing Description know the cloud is bound to the specific version of the TD. 
+
+Mikkel: We call this configuring the device. In the DB we change its representation. In modern devices everything changes. And they can be reconfigure them. Our TDs change. 
+
+Ganesh: thanks.
+
+Luca R: Hi thank you for the presentation. I saw you are in the progress on publishing, when is going to be published.
+
+Mikkel: Next year the draft will be voted upon in the ISO system. A couple of years should take. The iso working group only started a month ago. Experts from 12 countries. Very active group. We inherited a mature standard already. I don't remember the steps needed.
+
+Luca R: thank you.
+
+Ege: you mentioned the ontology should be good quality, do you have any recommendation for us? what is important?
+
+Mikkel: core is important. We have looked in IoT standards (SSN) also QUDT and OM2, we reference both of them. People can decide which one to use. Our favorite is QUDT even more if you think that it could be connected with EClass.
+
+Ege: we are working on an RDF version of EClass.
+
+Mikkel: If the twos (QUDT and EClass) can be joined that would be fantastic.
+
+Ege: you mentioned modbus, do you use any other protocol?
+
+Mikkel: BACnet, modbus, OPCUA (all ecosystem, seems like a clan). We have our own proprietary bus. The flexibility of WoT is great to add your own protocol. It makes a lot of sense.
+
+Ege: Nice to hear. Btw BACnet is now supported by WoT.
+
+Ganesh: One other fear is breaking changes that this evolving ontologies expose us to. Any ideas how to make the whole adoption more robust? What if the ontology needs to be updated? you simple go there and update the firmware?
+
+Mikkel: it is hard. In the essence the ontology is part of the software. In our internal ecosystem we have the concept of releases of ontologies. We download it and use fixed anchor to it. We use sematic versioning. I don't have silver bullet.
+
+Ege: good practice in general.
+
+Mikkel: we keep track of the software version in our model. We also have responsibility on how to support hardware parts. It is the same with the software. We have to keep things up and running for the product lifetime. Even contractual obligations. There is also data quality issues, what if the firmware version was wrong simply because it was not really important. 
+
+Ege: Do you have other examples of properties you would like to have. 
+
+Mikkel: Sometimes we have cleaver packaging of data. How to model this kind of objects in the TD? It can be complex as you wish.
+
+Ege: actually perfect timing. We discussed this issues. Felix is working on this too. Current point is to use pointers in conjunction with data schema. It is needed. The post processing is not happing yet.
+
+Michael Koster: Best thing you can do is using array schema.
+
+Cris: it is good to collect use cases to help the standard move on.
+
+Mikkel: storing data and do a little bit of post processing, we are starting looking at to math markup languages.
+
+Ege: Math json could be a starting point.
+
+Luca B: The use case is part of the whole data mapping concept that I am trying to put in wot 2.0 . I care more about extracting the date but today I learned that there is also the need to do some transformation (and/or filtering). We can use JSON Pointers to extract, jsonpath can do some operation on the data. JQ and Jack can do more manipulation, but they are not standard.
+
+Ege: W3C has a liaison with ISO we can have a preferred path to contribute to this use cases.
+
+Mikkel: plugfest are important to show the power of WoT rather than PDFs.
+
+## :envelope_with_arrow: Feedbacks
+
+- The data mapping use case is present. Mapping a json schema object to a modbus bit map.
+
+- Dynamicity of TDs. (and TD versioning and its impact on consumers)
+
+- We should have good quality ontologies in WoT.
+
+- A mathematical markup language is needed. Mikkel also says that putting lambda functions is a big security risk
+
+- Plugfests are important to show that it works. Mikkel's main selling point is actually this.
+
+## Presentation Summary
+
+* **Introduction:**
+  * Michael from Grundfos, Chief Data Scientist in data and AI, with a background in physics and mathematics.
+  * Member of ISO standards in Denmark.
+
+* **ISO Industrial Data Ontology (IDO):**
+  * Part three of ISO TC 184 Subcommittee 4 WG 26, focused on industrial data ontology.
+  * Represents industrial data, builds vocabularies, manages asset models using reference data libraries and reasoning.
+  * Derived from ISO 15926, part 14, now ISO IDO, used in various industries.
+
+* **Use Case: IoT Operating Instructions for Water Solutions:**
+  * Use case involves IoT operating instructions for water pumps and irrigation.
+  * Challenges in managing IT equipment for intelligent water solutions.
+
+* **Internal Data Models and Web of Things (WoT):**
+  * Addressing challenges through internal data models inspired by WoT.
+  * Aim to improve information handover between teams, departments, and companies.
+
+* **One-Slide Summary:**
+  * Many projects aim for intelligent solutions but struggle with information handover to cloud solutions.
+  * Utilizing open standards, such as WoT, facilitates effective communication between embedded software and cloud solutions.

Presentations/2023/Readme.md

@@ -5,8 +5,10 @@ If you have online versions, you can link them here in reverse chronological ord
 
 ## Online Presentation Resources
 
+- W3C WoT Meetup 11 - ISO Industrial Data Ontology - Grundfos | [Video](https://youtu.be/VYHjMU5a9FI) | [PDF](./2023-12-WoTCG-Meetup11-MikkelHBrynildsen-ISO_IDO.pdf)
+- [W3C WoT Meetup 11 - Intro](https://docs.google.com/presentation/d/1r_7PrstswVoU_1awFvrcfa6tuQRzm76coDZkPfuY68k/edit?usp=sharing) | [PDF](./2023-12-WoTCG-Meetup11-AguzziKorkan.pdf)
 - W3C WoT Meetup 10 - Designing an API for the Physical World - Seam | [Video](https://www.youtube.com/watch?v=LHybkrb3Xek) | [PDF](./2023-10-WoTCG-Meetup10-Seam-SyBohy.pdf)
-- [W3C WoT Meetup 10 - Intro]([https://docs.google.com/presentation/d/1j1trj42kzy7FWaYR3xne39Ic_PJqjszjEmXIww7x_qU/edit?usp=sharing](https://docs.google.com/presentation/d/17Whb-1nKF949bWN3dXjEin7q0nuJW9kUsVXgp3HlcLY/edit?usp=sharing)) | [PDF](./2023-10-WoTCG-Meetup10-AguzziKorkan.pdf)
+- [W3C WoT Meetup 10 - Intro](https://docs.google.com/presentation/d/17Whb-1nKF949bWN3dXjEin7q0nuJW9kUsVXgp3HlcLY/edit?usp=sharing) | [PDF](./2023-10-WoTCG-Meetup10-AguzziKorkan.pdf)
 - W3C WoT Meetup 8 - Advantages and Challenges of Web of Things for Digital Twins - Schaeffler | [Video](https://www.youtube.com/watch?v=kPfdqGYvBVM) | [PDF](./2023-05-WoTCG-Meetup8-Kuestner_Schaeffler.pdf)
 - [W3C WoT Meetup 8 - Intro](https://docs.google.com/presentation/d/1j1trj42kzy7FWaYR3xne39Ic_PJqjszjEmXIww7x_qU/edit?usp=sharing) | [PDF](./2023-05-WoTCG-Meetup8-AguzziKorkan.pdf)
 - W3C WoT Meetup 7 - Web Scale Agents with WoT and No-Code Applications in IoT - IntellIoT, Simon Mayer | [Video](https://www.youtube.com/watch?v=iqNX9DgFaBM) | [PDF](./2023-04-WoTCG-Meetup7-Mayer-IntellIoT.pdf)