SolidOS Team Meeting

# SolidOS Team Meeting * Date: 02.11.2022 18:00 (UTC+1) * Call: [https://meet.jit.si/solid-operating-system](https://meet.jit.si/solid-operating-system) * [Previous meeting link]() * [Next meeting link]() * Chat: [https://gitter.im/solidos/solidos](https://gitter.im/solidos/solidos) * Repository: [https://github.com/solidos/solidos](https://github.com/solidos/solidos) * Meetings home: [https://solidos.solidcommunity.net/public/SolidOS%20team%20meetings/](https://solidos.solidcommunity.net/public/SolidOS%20team%20meetings/) ---- ## Present attendees * Jon, Angelo, Noel, Matthias, Alain, Timea, Tim, Rahul ### Scribes * Matthias and Timea --- ## Topics ### Round Table #### Alain * I rewrote context-pane hoepfully resolving the issue. I tested it well and I do not see any problem. * the new datamodel is not compatibal with teh new one we need to discuss a potential migration. #### Timea * all my free time I used keeping up with WebID spec. * no further updates. #### Jon * had an online meetup with 4 people who showed up * we scheduled an inperson meetup on the 29th Nov. #### Matthias * I checked the server before starting the meeting * 2.5k users on solidweb.org * 102 on solidweb.me * I plan to do a data back-up every week slowly I feel responsible for them and I want to do it right. * Alain: I made a bechul on solidcommunity.net so I can help you there. #### Noel * last week was done with new features. Now I can start to document. I will try to do the basics and get them out the door. * I was looking at history, I started ldeveloping Dec 2 years ago. I want to release it before 14th Dec so it is exactly 2 years :) * Been talking about the BBC Solid app in the forum: https://forum.solidproject.org/t/implementation-of-bbc-together-data-pod/5763 #### Tim * I was in Texas for the Austin Film Festival and I remember we used to have someone who wanted to starta Solid chapter there, but don't remember who. * On Friday, John and I talk at the WebSummit in Lisbon Portugal. * has everyone seen the BBC video? Check forum for link. Test it and give them feedback if you find something. It is not reachable if you are outside UK. * did a keynote in a Singapore event this morning, online. * thanks Alain about the fix, we can talk about migration. #### Rahul * I'm in the community for about 1 year now form notifications. * I was working on syntropize -> see my intro from last week. Find links and motivation in last meeting. * I was helped a bit by Jeff (thanks to Jeff and Sarven) #### Angelo * currently working on a databrowser with my components. I linked my Pod but I need to add more components to make it more useful. * one can navigate through the data already ### Technical topics #### Migration issue with the data model * Alain: the problem is when we have a context-pane and the data model has changed. If you want to use the old info - there are different ways to do that. * 1st problem -> there is no version saying whcih datamodel is used. I do not know if we need to create it or we have to create a new pane like it was done with long-chat and short-chat. * idea: create a migration function because in the cintacts-pane there are tools and it is possible to add tools to create a migration. I am not sure if it is reversible. If you do the migration, not sure if you can redo it without any problem. * PROPOSAL: not to do anything and to do it manually. Maybe there are not a lot of context data. I am open to all ideas. * Noel: who changed it and why? * Alain: because I suppose... I am the core of the problem. I tried to use reasoning to find a solution. it is working well with NSS but not in CSS. if we ant to use more solid servers that do not accept reasoning we need to have a non reasoning solution. I implemented a non resoning solution. It works apparetnly quite well but it is not compatible. * Noel: what do you mean by reasoning? I understand this to be on the client - equivalences in vocabulary. * Tim: RDFLIB had a switch turned on on NSS which can reason over an array of thinks (A sameAs B for example and a bit of rdfs (inverse property)). When you do a query, it will use the canonical one and it is more powerful with reasoniing. In Solid one should be careful abotu what you trust. CSS people is very clear they do not want WAC to propagate. So therefore the only way to use a group with WebID that the WebID used for authentication is in the group member directly. And not follow the sameAs. What it doe snot do is to go over all the entries on the grouop file and fix them. With the tarcker I had such a issues. You had a order list of states and a rdf:list of states - to make sure it is the same. If they are different it patches the config file to match. * Tim idea 1: put up an alert - there is a config issue - should I fix it? Yes or No. * Tim idea 2: Or have a sanity check on each group. This group has issue - is it ok if we clear up issue? Context have a whole kind of tools to clean up duplicates for example. * Tim: add a generic tool that does it but asked before 'are oyu sure?' * Noel: the data itself there is not right or wrong. with the new one it will work but with the old not. People will only care if they use old apps. * Tim: they care if tehy use the group inan access controll - that breaks. * Neol: but we could fix it as well, right? We can fix ACLs too or? * Tim: we are not changing teh ACls only the group card the group list that is used for teh ACL. * Angelo: thsi is exactly what I did for me so that the froup works for teh ACL. * Tim: we need to document this under contacts! this is the most improtant in contacts! * Alain: when you access teh data with different mashlib - old versions will not work. * Tim: really? because they do not do sameAs * Alain: old mashlib will not work. Reason why I speak of version. If you do not know the version you cannot have somehting that is working. There is no way to have a pane with a data version system. * Tim: we could make 1) I thought we had a spec for this somewhere - I though othe rpeople asked how to implement contacts... or at least shapes? * Alain: yes we have a shape. * Tim: could we fix it by doing a ACL shape validation using SHACL shape? * Alain: I have no idea * Tim: I bet we could. When I converted my contacts I made a shape for the context and converted it from the Mac. I did a validation to make sure the import was done good. Checked for people with no name or only one name... * Alain: 1) new version ??? * on the long run new version id data model changes. * todo create an issue on the contact pane. * Alain trying on the test server * Tim: it will only work with actual mashlib * maybe flag the version * good to be backwards compatible * Alain: conclustion -> we will try to publish a mashlib with new profile-pane and context-pane and there will be an issue to discuss data model. I will try to see if we can make mashlib able to read old context-pane (backward-compatibility). #### Rahul demo * it is more about the way the system works. * shares his screen - we can think of it as an App but it is intended to be treated like an operating environment (how Windows was for MS DOS, up until v3.1) * Timea observations: * views are like our panels, Cards (which are in plain JSON) are like elements in teh dashboard maybe. * cards are very flexible in what they can display (from web or from Pod) * views of context is done dynamically - keep the order of data. * each card is a folder on the Pod that loads the json * it is fast for publishing * for viewing ddata -> you input the url and the type of view. The app will dynamically load data models as it needs. * Noel: it reminds me of Digital garden (https://nesslabs.com/digital-garden-set-up) - the idea you have different nodes and you can connect them to each other. * Write up on how the code loads data dynamically here: https://hackmd.io/EmVuvlkQTUWWvD9iLfhwIA * More docs at <syntropize.com/docs> #### ---- This template is based on the [W3C meeting template](https://github.com/solid/specification/blob/main/meetings/template.md) [Code of conduct](https://github.com/solid/process/blob/main/code-of-conduct.md)