Create 2025-05-14.md #731
Open
Create 2025-05-14.md #731
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
elf-pavlik
commented
May 14, 2025
| elf Pavlik: 20:46 | ||
| And I think before we agree on a solution, it's good to explore different paths because then you know, we hopefully get like the best of all worlds. So I think that's part of incubation, that we explore different way of doing something and then once we get more experience with them, say okay, now we understand the problem better possible solution. And hopefully with that experience we can arrive at a better final standard where we can reach consensus. So that's my approach practically. Okay, so that was site playground. Again, feel free to reach me on matrix or post comments somewhere like in a site repo, wherever you like. I welcome more feedback. Moving to the next topic. Link web storage use cases. I've seen some movement in the in the repo. I think there was first pull request merge with basic use cases. I understand Hadron is the the one leading the effort. Well, I remember Christoph at least you are also pretty involved in the use cases. I think also Eric is in a working group. Again I'm not following that. I am. Okay, so do you see what are the next step based on? Because it's like hundred something use cases. I also have seen requests from PR1 which is kind of like adding all of them, like just at least references. Do you know what's the plan for processing those use cases and integrating further in the document? | ||
|
|
||
| Christoph Braun (KIT) uvdsl: 22:25 |
There was a problem hiding this comment.
Suggested change
| Christoph Braun (KIT) uvdsl: 22:25 | |
| Erich Bremer: 22:25 |
elf-pavlik
commented
May 14, 2025
|
|
||
| Linked Web Storage Use Cases and Variants (20:35 - 28:29) | ||
|
|
||
| Recent progress in Linked Web Storage use cases; Hadrian leading, Christoph Braun becoming more involved as a spec editor. |
There was a problem hiding this comment.
Suggested change
| Recent progress in Linked Web Storage use cases; Hadrian leading, Christoph Braun becoming more involved as a spec editor. | |
| Recent progress in Linked Web Storage use cases; Hadrian leading, Erich Bremer becoming more involved as a spec editor. |
elf-pavlik
commented
May 14, 2025
| elf Pavlik: 29:21 | ||
| And we had discussion with Jeff about the catalog and also like at some point we arrived in the direction that what's in the catalog, it's just like a aggregation of things that are pretty much, you know, its own data domains and they will need more detailed work to, to be able to have apps to work with those domains. And there are already few events. So I think it's interesting angle that as we work on adding things to the catalog we can at the same time start working processing those data domains. My question here is how we want to continue with that. I know that we discussed using the solid shapes repo and maybe work on some shapes for events over there. There are a few people who did some work on calendars also noticed on the chat Melvin was saying he developed recently something there's Tim from Practitioners even though I don't think I've seen him seen him recently on on on the meetings and yeah. So do does anyone have some ideas how you would if you are interested first and secondary how you would like to see the next steps to like start moving towards some kind of, you know, mini speckle at least some basic shapes for calendars and events. Go ahead Jesse. | ||
|
|
||
| Marc Haddle: 30:45 |
There was a problem hiding this comment.
Suggested change
| Marc Haddle: 30:45 | |
| Jesse Wright: 30:45 |
elf-pavlik
commented
May 14, 2025
| elf Pavlik: 31:01 | ||
| Yes, I think the most established standard is I calendar that it's One of the RFCs I had linked last week. It's in a minute and I think that there's also RDF cal which I understand it's like mapping to rdf so that will be probably my first direction. So it's easy to consume the ICAL feeds and events which I think all of us use in our day to day coordination and have an easy way of importing them and also whatever we have in the storage to have a way to transform and export it to iCal. So that would be my preference to take it from this angle and as soon as possible bridge maybe the solid CG calendar make it available as RDF and as we start scheduling meetings with one another we also start eventually maybe in few months using the solid and still be able to export it to our Google Calendars, Outlook, whatever people are using. | ||
|
|
||
| Marc Haddle: 32:02 |
There was a problem hiding this comment.
Suggested change
| Marc Haddle: 32:02 | |
| Jesse Wright: 32:02 |
elf-pavlik
commented
May 14, 2025
| elf Pavlik: 33:22 | ||
| Okay, awesome, we'll take a look at it. And again I also invite people tomorrow during practitioners meeting, whoever is interested to explore it together. And yeah, I'll try to, if not every week, like every other week to have some update and see if we can make stable progress on that. By the way, I would ask everyone I mentioned that we are testing again the fireflies to scribe, but please add yourself to the present so at least we do that manually for now. Okay. Are there any other comments on that topic? Okay, hearing none. We can move to another one and then either we have like any other topics or we can finish earlier. So this one is related to some prior conversations. I had a few questions to Joachim from CSS a few months ago and so the next topic it's a solidcomputer net bring your own custom domain and I know that CSS does not support having custom domains for ports. I understood from my go ahead it. | ||
|
|
||
| Marc Haddle: 34:51 |
There was a problem hiding this comment.
Suggested change
| Marc Haddle: 34:51 | |
| Jesse Wright: 34:51 |
elf-pavlik
commented
May 14, 2025
| Theo @thhck: 46:16 | ||
| And take the temperature and then maybe we can have a meeting altogether. | ||
|
|
||
| Marc Haddle: 46:20 |
There was a problem hiding this comment.
Suggested change
| Marc Haddle: 46:20 | |
| Jesse Wright: 46:20 |
elf-pavlik
commented
May 14, 2025
| ### Scribes | ||
| * https://fireflies.ai/ | ||
|
|
||
| --- |
There was a problem hiding this comment.
Suggested change
| --- | |
| ### Links | |
| * https://rallly.co/invite/3y8dd4bwJ56C | |
| * https://github.com/w3c-cg/awesome-semantic-shapes?tab=readme-ov-file#shape-convertors-and-generators | |
| * https://shacl-play.sparna.fr/play/convert | |
| --- |
uvdsl
suggested changes
May 14, 2025
There was a problem hiding this comment.
I think the AI-generated minutes are okay-ish at best. It is really not great that the AI wrongly attributes full statements to the wrong person. Also, when multiple persons are speaking, the AI is not able to distinguish by voice. Kindly consider suggested changes. Cheers!
| Meeting Scheduling Issues: Low attendance linked to current meeting time; a scheduling poll is created to optimize weekly meeting slot. Manual attendance required while AI tool is being tested for notes. | ||
| SAI Playground Demo: Elf Pavlik showcased early SAI playground version for visualizing data grants and delegation. Feedback on usability and clarity for authorization flows requested; emphasis on improving representation of delegation chains. | ||
| Delegation & Policy Enforcement: Discussed access grants vs. third-party sharing limits. Need for various implementation approaches before standardization. Agreed on necessity of technical enforcement and potential legal frameworks. | ||
| Linked Web Storage Progress: Led by Hadrian with Christoph Braun as a spec editor. Preference for adding use case variants to existing scenarios rather than creating new ones. Suggestions to document subtleties via GitHub. |
There was a problem hiding this comment.
Suggested change
| Linked Web Storage Progress: Led by Hadrian with Christoph Braun as a spec editor. Preference for adding use case variants to existing scenarios rather than creating new ones. Suggestions to document subtleties via GitHub. | |
| Linked Web Storage Progress: Led by Hadrian with Erich Bremer getting more involved, potentially as a spec editor. Preference for adding use case variants to existing scenarios rather than creating new ones. Suggestions to document subtleties via GitHub. |
| Send the finalized scheduling poll to the mailing list and invite wider participation (06:06) | ||
| Link the GitHub issue regarding delegation limitations and share with Christoph Braun (16:09) | ||
| Contribute use case variants or subtleties to the Linked Web Storage use cases, either by commenting on existing issues or linking related issues (28:29) | ||
| Coordinate with Hadrian and others to propose a project board or systematic way to track community development priorities and funding opportunities (48:10) |
There was a problem hiding this comment.
Suggested change
| Coordinate with Hadrian and others to propose a project board or systematic way to track community development priorities and funding opportunities (48:10) | |
| Coordinate with Hadrian and others, including Jesse and the ODI, to propose a project board or systematic way to track community development priorities and funding opportunities (48:10) |
| Oh, for tomorrow? Awesome. Yep. He didn't announce it I think but it would be nice surprise. | ||
|
|
||
| Christoph Braun (KIT) uvdsl: 04:04 | ||
| Yeah, I think he. He put me on the schedule like that was his message from. Let me double check. Yes, I'm scheduled for May 15, 15 UTC. |
There was a problem hiding this comment.
Suggested change
| Yeah, I think he. He put me on the schedule like that was his message from. Let me double check. Yes, I'm scheduled for May 15, 15 UTC. | |
| Yeah, I think he put me on the schedule like that was his message. Let me double check. Yes, I'm scheduled for May 15, 15 UTC. |
| Okay, so it's mostly to help like navigate how the delegation works, how the data grant work. And the main assumption inside is that you have multiple autonomous agents which can be individuals or organizations. We call them social agents to distinguish from applications which are software agents. Each social agent here, they act as a data owners or resource owners can have multiple storages. So as we see, acme's organization has two ALICE has two storages and BOB has one. Again, it's up to everyone to choose how they organize the data. But it's assumed that people can have more than one storage they own. And so in those kind of like columns we have those data owners with their storages. And now we can select the end user, so the user who will be accessing the data. So for example, if we select alice, we can see that there are a few data grants here. Here, for example, ACME granted ALICE access to some data in this storage. I will add like the, the indication of which shape trees. So actually in the storage, the way I implemented it, we have things kind of partitioned into data registrations. So there are visualized here like calendars, events, videos as an examples. So also those data grants will be matching specific data registrations. So here just kind of shows that maybe I'll turn off this one for now that the data owners granted access to specific social agents. But also besides social agents, we have applications. And that's also very important inside. And that's like the basic case for delegation. So now we can see, for example, when we click on this data grant, okay, you can see that this one is just giving access to one data instance. It's just highlighting here if we click on this one, okay, I have some issues. Okay, it gives access to the whole registration because the scope is all. But now we can also select application. For example, ALICE didn't authorize just application to access of ACME data. What we can see here based on the data grant from Bob, this is granted from Bob as a data owner. So again the columns are. I have some problem when I share screen. It's really slowing down. So again the columns are the data owners. So you can see that Bob granted access to ALICE and that's the whole registration of calendars and then ALICE delegated to the application she's using that only it can access some selected ones. And you can see you can only select access those selected two. And the last I'm wrapping up this demo and again it will look better just like placeholders for now. And then if we look for example of Bob, this is a little more complex because we have Bob as an end user. So these are the grants that are issued for Bob and Bob is using this application. So here we have a longer delegation chain because as you see the original grant, ACME granted access to ALICE and then ALICE delegated that access to acme's data. Again, what's in this column? It's mapping to this storage. If there was something to this other storage, there would be a second column here. So this data for this data registration in this storage, ACME granted access to all the calendars to alice. Then I'll of course. Okay, we can click on it. You will see that it's all the calendars you can see highlighted in green. Then ALI is delegated only selected ones to Bob and you can see there's only one. And then Bob also delegated the same selected ones to the application he's using. So this is like actually a little more involved delegation that we have. The data owner is acne to Alice, Alice to Bob, and then Bob to application. And we always assume that the end of the Chinese application because I don't know, maybe when we get some neural links we can directly access the data with our brains. But most of the time we use some kind of client application. So there's always like the, the final delegation, the tail of the delegation chain is the application. So that's pretty much what I have so far. There's plenty things I have to add, like the inherited data grants to show you can access calendars and events that are kind of related to those calendars. There are a few other things. Access authorizations to show that you can set a policy that it's spanning across multiple storages and even multiple agents and their storages. So access authorizations give you like a higher level of policies that you can set that it's not limited to a single storage. So there are quite few things. My main kind of request for feedback. If you looked at SAI and you find some things hard to understand and you think that this kind of way of visualizing it could be like, you know, interactive way, useful to understand it. Please let me know like what you would like to see that you can, you know, click around and select some things and, and hopefully it would help people understand those nuances. I know that some people find it complicated. I think once, once kind of it's explained well, it's probably easier to, to understand than just by reading the spec. So there are Already two primers. And this would be like a. Another accompanying resource to. Yeah, when we have conversations or you know, someone tries to explain something, eventually one should be able to load it and you know, maybe share screen and then as we discuss select something and say okay, that's how it works that it's how it's intended to work. This was my little demo here, let me now turn it off and I will leave from the other browser. Any questions, suggestions on that topic? Go ahead. Christoph. | ||
|
|
||
| Christoph Braun (KIT) uvdsl: 15:01 | ||
| Yeah, thank you, Pavlik. So I think the visualization definitely will. |
There was a problem hiding this comment.
Suggested change
| Yeah, thank you, Pavlik. So I think the visualization definitely will. | |
| Yeah, thank you, Pavlik. So I think the visualization definitely will help people to understand the topic. |
| Help. | ||
|
|
||
| Christoph Braun (KIT) uvdsl: 15:11 | ||
| Understanding how like the. Which resources in particular are given access. I think the. From a visual standpoint, maybe you could manage to make it a bit clearer that there is this chain of authorizations, that these really depend on each other. This was the one thing I was a bit missing from the status quo. But again I think it's a good, good aid definitely. I'd have a question regarding the delegation process. Can a. If, if an access grant or an authorization access authorization is given, can a party decide not to enable third party sharing? So like Acme disables that Alice can delegate to Bob or whatever. Did you think of that? |
There was a problem hiding this comment.
Suggested change
| Understanding how like the. Which resources in particular are given access. I think the. From a visual standpoint, maybe you could manage to make it a bit clearer that there is this chain of authorizations, that these really depend on each other. This was the one thing I was a bit missing from the status quo. But again I think it's a good, good aid definitely. I'd have a question regarding the delegation process. Can a. If, if an access grant or an authorization access authorization is given, can a party decide not to enable third party sharing? So like Acme disables that Alice can delegate to Bob or whatever. Did you think of that? | |
| I think from a visual standpoint, maybe you could manage to make it a bit clearer that there is this chain of authorizations, that these really depend on each other. This was the one thing I was a bit missing from the status quo. But again I think it's a good, good aid definitely. I'd have a question regarding the delegation process. If an access grant or an authorization access authorization is given, can a party decide not to enable third party sharing? So like Acme disables that Alice can delegate to Bob or whatever. Did you think of that? |
| Go ahead, Chris. Go ahead, Christoph. | ||
|
|
||
| Christoph Braun (KIT) uvdsl: 25:04 | ||
| Yeah, I agree with that. And I, so I'd agree with that. And it makes the subtleties apparent or more obvious because I think some people from the community, they're, they're quite clear on that we need this distinction between the authorization app or agent or whatever mechanism there is to, to authorize access, that this should be distinct from some consuming business application or personal application. Right. So there, there is this notion that people are aware of and I believe these people, me included, will like fight for that. But still, if there are any use cases you can give us as ammunition, so to speak, that would be great. |
There was a problem hiding this comment.
Suggested change
| Yeah, I agree with that. And I, so I'd agree with that. And it makes the subtleties apparent or more obvious because I think some people from the community, they're, they're quite clear on that we need this distinction between the authorization app or agent or whatever mechanism there is to, to authorize access, that this should be distinct from some consuming business application or personal application. Right. So there, there is this notion that people are aware of and I believe these people, me included, will like fight for that. But still, if there are any use cases you can give us as ammunition, so to speak, that would be great. | |
| Yeah, I agree with that. And it makes the subtleties apparent or more obvious because I think some people from the community, they're quite clear on that we need this distinction between the authorization app or agent or whatever mechanism there is to authorize access, that this should be distinct from some consuming business application or personal application. Right. So there, there is this notion that people are aware of and I believe these people, me included, will like fight for that. But still, if there are any use cases you can give us as ammunition, so to speak, that would be great. |
| Yeah. So my main dilemma here is that I would want to introduce completely new use case, building a new scenario or maybe work with variants to say, okay, we have the existing use case, let's stick with that, but add a variant to it to say, but it's the same situation, but there's additional kind of requirement of saying actually Alice, don't want to just. Or whatever. Ginion or seven of hearts, whatever names we are using that they don't want to disclose the whole social graft application. So instead of introducing completely new use case, my preference would be just to have a use case and then like some variants of those subtleties, but not like completely separate use cases that someone has to imagine the whole scenario from scratch. And then maybe because of those differences in the scenarios, like different names, different whatever, we would like get distracted from what actually differs in those different variants. And for me would be important how we can actually work with those variants. Say, okay, the whole thing is the same, but there's one variation on that. | ||
|
|
||
| Christoph Braun (KIT) uvdsl: 27:13 | ||
| I mean from just an operational standpoint, either comment on the existing issue or if you can. I'm not sure how that works on GitHub actually I've never used that feature. Open a sub issue or create a new issue, link the old issue, say everything is the same except for these things. I think then that would also work except Eric has a different opinion on that. I'm sorry, follow your point. So I just was listing ways that would accommodate like public's idea of reusing an existing use case as a basis and then introducing subtleties or like more details, more restrictions on the use case. And I was just saying, well, link the old, link the original use case or comment in that use case with that variant and then people will be able to find that variant and pick up the requirements from there. That was my suggestion. Yeah, I, I would agree with that. |
There was a problem hiding this comment.
Suggested change
| I mean from just an operational standpoint, either comment on the existing issue or if you can. I'm not sure how that works on GitHub actually I've never used that feature. Open a sub issue or create a new issue, link the old issue, say everything is the same except for these things. I think then that would also work except Eric has a different opinion on that. I'm sorry, follow your point. So I just was listing ways that would accommodate like public's idea of reusing an existing use case as a basis and then introducing subtleties or like more details, more restrictions on the use case. And I was just saying, well, link the old, link the original use case or comment in that use case with that variant and then people will be able to find that variant and pick up the requirements from there. That was my suggestion. Yeah, I, I would agree with that. | |
| I mean from just an operational standpoint, either comment on the existing issue or if you can. I'm not sure how that works on GitHub actually I've never used that feature: open a sub issue. Or create a new issue, link the old issue, say everything is the same except for these things. I think then that would also work except Erich has a different opinion on that. | |
| Erich Bremer: | |
| I'm sorry, follow your point? | |
| Christoph Braun (KIT) - uvdsl: | |
| So I just was listing ways that would accommodate like public's idea of reusing an existing use case as a basis and then introducing subtleties or like more details, more restrictions on the use case. And I was just saying, well, link the old, link the original use case or comment in that use case with that variant and then people will be able to find that variant and pick up the requirements from there. That was my suggestion. | |
| Erich Bremer: | |
| Yeah, I would agree with that. |
| elf Pavlik: 36:39 | ||
| Your comments when it comes to hosting it yourself. So for example, for most of us, Probit don't not a problem to you know, spin up a virtual machine somewhere VPS and maybe, you know, deploy it with Docker or Kubernetes or whatever people are using. For example, if I take an example of my wife, she has a domain name, I think she uses something like wix where there's a simple instruction like hey, go to your domain register, add this to DNS and then we handle the rest for you. And I think it's a bit a different step for a person to just configure custom domain with their domain name registrar and run their own machines, maybe deal with the backups and if something goes wrong, how to fix it. So I think there is a gap there when it comes to solid community.net I agree that if someone wants to have a more professional hosting, maybe they should use a different host. This is more like a demo. But at the same time I think there's a use case here that if someone wants to just get started with Solidcommunity.net they have the domain name, they can configure it and then they can have a choice of saying hey, actually you start outgrowing this greenhouse incubation account that we offer as a courtesy and maybe you want to move to more professional provider with whatever needs, maybe bigger quotas and so on that you have. And then if you use your product custom domain from the beginning, transplanting you into like a bigger pot, it's a pod, it's not a problem anymore. So I think it actually could be beneficial even for solid community.net if there's no intention of having more like a professional hosting to have people use their domains if they have it, because then it's easy to tell them like hey, please just find a different house. You, you already kind of like grown up, move out and, and that's my idea. But either way I think it would be great if CSS supported that. So for example, I know what others think, but maybe you know, we can, if we agree that's a good functionality to have, maybe we can like, you know, look for the people who already have experience working with CSS and you know, look for a project or funding or something that someone could add this functionality. And I know that for example, Hadrian is a big proponent of using something completely different like this or something like that, that is a completely different resolution mechanism and you kind of become independent from the domain name. But as I was saying earlier, I think we should explore different possibilities at the same time without putting all the baskets in eggs in one basket. And this is like very common accessible approach that you know, for many I don't. I'm not saying that it's everyone you know, it's capable of having their domain name, but many people nowadays, especially the peoples who I would consider early adopters of solid people who, you know, use Fediverse, maybe use Indie web or something, it's probably very accessible for them. So I think if we have already capability of saying hey, bring your own domain name and then if you move, just move with your domain it's easier this way. I think this is benefit but I would like to hear other people perspectives and Jesse, please go ahead. First. | ||
|
|
||
| Marc Haddle: 39:56 |
There was a problem hiding this comment.
Suggested change
| Marc Haddle: 39:56 | |
| Jesse Wright: 39:56 |
| Theo @thhck: 45:08 | ||
| Sure, yeah. It was that we can make a component to CSS to make people use their own domain name, but we can also work at the hostel level. And that reminds me that at last post them I meet like in the hoster. I think. | ||
|
|
||
| Marc Haddle: 45:36 |
There was a problem hiding this comment.
Suggested change
| Marc Haddle: 45:36 | |
| Jesse Wright: 45:36 |
| I agree. But if there's something we would like to coordinate and maybe even have like a clear way of communicating, we publish a project board and then in your grant you can link to it say hey, this is something that it's really demanded by the community. That's like a way to show it. I think it would be like a plus for your proposal that you could bucket with something saying hey, it's not just my personal idea, I think it would be cool. But actually it's something that you know, there's a clear signal from the solid community that many people would like to have this functionality available. So it's. It's mostly saying if in those two weeks, we can move a little faster than usual and. And have something that you can use as a kind of reference to say, hey, this is something that can back your proposal and hopefully help you secure funding to work on it. Okay, we have a few minutes left. Are there any final remarks, questions, suggestions, proposal for the next week agenda? I know that Hadrian will be preparing it, but the part is available here. But also, if you have an idea to just throw out, I will be happy to coordinate with Hadrian and propose it for next week. Okay, so reminder, tomorrow we have a solid practitioners meeting. It's in the CG calendar. If you subscribed, you should see it in your calendar. I'm looking to Christoph's presentation about delegation. This is super exciting to me. And with that in mind, I wish you all good evening, afternoon, whatever time zone you are at, and hopefully see you tomorrow. And if not, next week. | ||
|
|
||
| Christoph Braun (KIT) uvdsl: 53:05 | ||
| Thank you, Pac. All right, bye, everybody. |
There was a problem hiding this comment.
Suggested change
| Thank you, Pac. All right, bye, everybody. | |
| Thank you, Pavlik. All right, bye, everybody. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
I think AI did a decent job and everyone was able to fully participate.
Some glitches seem related to me not muting my mike and confusing the bot with activity on my audio feed. This should be solved if we all make sure to mute whenever we don't speak.
We may need to create a script to format it better in the future.