Meeting minutes
<pchampin> not easy for me to speak
<pchampin> am I sill muted on zoom?
<laurens> yes
<pchampin> for future scribes, instructions are available here: https://
<bumblefudge> present?
Introductions
<csarven> Hi I'm Sarven Capadisli. My WebID: https://
csarven: hi, I'm Sarven, excited to be here, it's been a long road
… I've been involved in Solid (core aspect of Linked Web Storage work) since 2015 at MIT
… I've chaired the Solid CG for 4 years, authored and edited almost all of the input docs you see in the charter
… and worked on implementing applications and some server work
… happy that a lot of fun and pain and crying led to the work here, hoping we'll have a successful outcome (and I'm sure we will!). and fix the web!
acoburn: hi I'm Aaron Coburn, work for Inrupt, I'm one of the co-chairs
… been working in the Solid space for a number of years, worked on LDP related things, including Fedora API, I was editor and author on that spec
… very happy to be here, I've also been editor and author of various docs in the Solid space as well
eBremer: hi I'm Eric Bremer, working for Stony Brook University in the Department of Biomedical Informatics
… working on deep learning pipelines, GeoSPARQL, modeling the storage system after Solid
laurens: hi, I'm Laurens, one of the co-chairs, I represent the government of Flanders,
… been working with Solid for many years, mostly apps building on top of it
… I hope for a very successful outcome for this WG of course
dmitriz: hi, I'm Dmitri, longtime fan of Solid
bumblefudge: hi, I represent the IPFS Foundation
jeswr: I'm Jesse Wright, been working on Solid and semantic things for several years, previously affiliated Australian nat'l university
… worked at Inrupt a couple years working on SDKs, now doing research on semantic web agents, data privacy, automated data governance
hadrian: hello I'm Hadrian, excited to be here, I'm representing Inrupt, I'm also co-chair of Solid CG
… also member of Apache Software Foundation. also happy this WG started
ericP: my name is Eric P., worked at W3C for a long time, on a number of semantic web standards
… my current affiliation is - Leiden (sp?) University Medical Center
… I found out today that I'll be doing another medical data storage system on top of Solid. I'm also co-chair
benDM: I'm from Imec, Belgian research university, we've been doing some research since 2009, me from 2013
… we've been doing Solid research and working together with Flemish Government for a couple of years, bringing Solid to production
jucanbe: hi everyone, I'm Juan, I work on University Polytechnical Madrid, finishing my PhD Thesis
… related to semantic web policy policies based on ODRL
timbl: hi, I'm Tim Berners-Lee, invented the web 30-something years ago,
… worked at MIT at W3C, which a bunch of you contributed to
… in 2001, we had some funding in the lab from DARPA for some semantic web research
… so we built rule-based system, exporters & importers, a Read-Write Web
… when browsers got Javascript / AJAX, the ability to communicate with the web itself, that meant we could move from apps running on our laptops to a read-write web
… if a storage system doesn't care which app writes to it, that's very powerful
… still working on a good name for it, we ended up with Social Linked Data or Solid
… we had more funding to improve Solid from Mastercard, in the lab
… and a few years, we founded Inrupt
… to continue the work of bringing Solid to the world
… my dayjob is at Inrupt
AZ: I'm Antoine Zimmermann, professor at Ecole Institut Mines-Télécom
… I've been doing research in the area of Semantic Web for 20 years now, recently I took interest in helping autonomous web agents to interact with web resources
… I'd like to investigate how artificial agents could have a web presence in the form of Solid pods
… also co-chair of the Web Agents CG
TallTed: hi, I'm Ted Thibodeau, involved with OpenLink Software since 2000, w3c shortly after that
… if it has to do with the semantic web, sparql, etc, I'm there
<pchampin> Pierre-Antoine Champin, from Inria and W3C. I'm the team contact of this WG. Thrilled to have the group starting at last, and to see you all here.
<AZ> s/at Ecole ...?/at École des mines de Saint-Étienne, part of Institut Mines-Télécom/
TallTed: my primary coding language is English, so, expect lots of grammar & spelling etc corrections on your PRs
<TallTed> :-)
ryey: hi I'm Rui Zhao, at University of ...?, working on privacy preserving computation and privacy policies
… big fan of decentralized tech, including Solid
<ryey> present
cpn: hi I'm Chris Needham, representing the BBC, I have various roles at w3c, media related groups
… the BBC has been very interested in personal data stores and a more private web in general, giving data ownership and control to people over the years
… we've had quite a few research projects in this space over the past 8 years. Relative newcomer to Solid myself
… as an org we're generally very supportive of the work here
ericP: next thing is to go over w3c Process a bit
ericP: W3C is a standards org that develops in the open
… and we make it so that the people in the WG can vote, argue, make PRs etc
… the public in general can also make comments at regular intervals (on Working Drafts)
… we'll probably start with a Use Case Document, then start working on specs after that
… the process is one where we as a group elect an Editor
… we make suggestions ,they make edits, we make a snapshot ready to share with the world (Working Draft)
… specifically First Public Working Draft (FPWD)
… after that, we'll iterate, publish WDs every 3 months so that the world is appraised
… we want the community engaged, making PRs
… for patent protection, everyone who joined the group had to sign the IPR policy
… goal there is to make sure there's no submarine patents
… if a member of the public makes a technical PR to a spec, we have to make sure to clear it
… then we'll go to Last Call. typically before Last Call, we'll make a Test Suite
… we'll have some specs, some tests, interoperable implementations
… the idea is separately implement it independently
… we're testing the spec by implementing. and then test the implementations
… then we get to go to Proposed Recommendation (PR)
… (this is also mostly for patent protection)
… there's another path, if we have a continuously changing spec, we can do what's known as an Evergreen Spec
… basically saying - it feels done, but we keep adding more cool things to it, so we can go to PR or REC when we do Evergreen as well
timbl: I'm sure if we missed any process aspects, we'll discover them and bring them up later
ericP: last thing is, during Candidate Rec (CR), when people are making comments, we have to document to the Director that we have adequately addressed all the issues people raised
… which means we have to do good bookkeeping, so, brace yourself.
<pchampin> just to point out that we have a list of groups with which we need to synchonize
<pchampin> we must not wait until going to PR to do that
ericP: yeah, another role of W3C is that we're telling a coherent story, that the groups talk to each other
… in the charter, we have a list of groups to coordinate with because they're related tech
… especially the Infrastructure group, Accessibiilty, etc
… so when we get close to CR, we'll have to do horizontal review in advance
… there is probably a fair amount of i18n and accessibility expertise in the group
… so if you have a question about that, please let us know
<TallTed> current W3C process doc -- https://
Charter and Resources
ericP: the charter has a bunch of pointers to the Mailing List and various input docs
<acoburn> Charter document: https://
ericP: the Process doc much more formally describes what I've been going over
… one of the things you'll see in it is "How do we know we're done?" or Success Criteria
… important to keep in mind, whether it goes into the critical path to completion of specs
… so, reading scope and deliverables is useful
<TallTed> all about LWS WG -- https://
ericP: also has a timeline
Vision
timbl: sure. the Vision for this WG is quite big, in that we're hoping that this protocol will be used by many systems
… if you have a pod which has one app running on it, that sort of misses the main value proposition
… but if you have different apps being able to read and write the same data, using read-write Linked Data Storage, and they actually share
… then that becomes very powerful
… obviously AI is part of the mix now, which adds to the power
… so, it's about putting the user in a situation where they're in control
… not just protecting privacy, but it's also focusing on the positive aspects -- individuals are a lot more powerful when there's interop between many different apps
… so, the vision is hard to explain to people. sort of like, Code Locally, Think Globally
ericP: any questions?
Objectives and Success Criteria
acoburn: (reads them from charter)
… couple of things we're looking at. one is, a proposed rec of interoperable implementations of each of the feature defined in the spec
… a Linked Data Storage protocol, defined at the end of this process
… we'll document the testing policies, processes, etc
… each spec will contain sections on Security and Privacy implications. there'll be a general thread model
… for the ecosystem
… in addition, each spec will contain a section on Accessibility
… benefits, impacts, accessible features
ericP: any questions on the above?
Testing
ericP: when we're developing the spec, we typically have 2+ implementations for each spec.
… there's a history of writing test suites for a protocol, notably LDP and SPARQL
… with protocol tests, it's more that - we have some client that we all use as a test tool, and it has some interactions with your server, it gets a response back, we validate if it passes/fails
… every feature in the spec needs 2+ implementations that pass the test
… so, we'll try to look to LDP test suite for inspiration
… so, expect that we'll be writing server tests
<pchampin> some work has also been done in the CG about test suites
Use Cases
acoburn: before we begin writing spec text, we're going to develop a Use Cases doc
… which will feed into a Use Cases and Requirements doc, which we'll prioritize
… and then we can buil da spec on that.
… we'll be collecting use cases, both from members of the WG and also from people who have interest in this
… there'll be a Github repo
<acoburn> https://
acoburn: two repos of relevance, you can find them from ^
<acoburn> w3c/
acoburn: in particular, the Use Cases Doc ^
… what we'd like to do is have use cases presented as Issues
… so, rather than PRs with specific use cases, instead open issues for each one
… that way, it can be refined, discussed, and once we have consensus, then we can make a PR
… which will eventually become a WG Note
… we don't have a template currently, but we will soon
… we began collecting some use cases from TPAC, at a breakout session
… so we have some already collected. those have not yet been added to the repo
… we'll open issues for them shortly.
… I suspect the use cases will primarily be collected asynchronously, and during meetings
… anything that requires specific discussion and voting, that's when it'll happen
<Zakim> pchampin, you wanted to talk about CR and exit criteria
benDM: the target goal of these Use Cases, is it meant that these are real world scenarios that we want to make sure works with LWS recs?
… are we talking about more technical use cases? What's the level of technicality required?
acoburn: I'm expecting things from the real world, things a developer or a user would encounter in the process of using this protocol
… so, as much details as is relevant to that. I don't want these use cases to be pages and pages in length
ericP: yeah, agree. typically Use Cases drive technical requirements, but they dont' initially contain technical reqs
… it's handy to have short coherent stories at the beginning
<Zakim> timbl, you wanted to mention the solid roadmap for vertiicals
<timbl> Solid project Verticals Roadmap https://
timbl: the Solid project has a Solid Pod provisioned for it
… there's some brainstorming for which Verticals are out there
… rough high-level list
… so, Healthcare community for example
… might have specific requirements.
ericP: yeah, we might need to make sure we've got coverage of use cases and verticals
… so, lots of eyes and brains on it
… one thing we can do is tag our usecases with which vertical it applies to
timbl: that would be good
… it'd be good if communities brought their existing specs and usecases, and we could map it to usecases
ericP: makes sense
acoburn: we haven't formally set a date/time for calls
… looks like Tue/Wed/Thu are very busy for everyone. so I suspect this will be a Fri or Mon call
<Zakim> csarven, you wanted to talk about degree of commitment to implement a requirement (derived from UCR) and share implementation code
csarven: just so we know when we're done (with use cases)
… we need to figure out a bunch of things - deadlines, some sort of criteria for done-ness
… criteria for which usecases to keep, or which ones are duplicates
… whether the usecases are ethically grounded, etc
… one thing I often find comes up with other groups is, assessing whether the usecases will actually be implemented, and by whom
… so that we're not overwhelmed by too many usecases / hypotheticals
ericP: makes sense
… there will be a lot of organic development, feel free to raise the issue if it doesnt get addressed
… we'll get to a point where we have use cases, where we'll be like - love to do that, won't get it done
ryey: also question about the scope of the usecases
… do we have a scope of what goes into the LWS protocol?
… in terms of what additional things go in there?
ericP: not yet
ryey: I have some use cases, but I'm not sure if they fit into the scope of the WG
ericP: write them up and bring em up in the meeting, or just open an issue
acoburn: we're going to need someone who can be an editor of these usecases. (one or more people)
<TallTed> for future -- meetings should target :55 as end-time (and really try hard to stick to it)
acoburn: to refine them, edit them, format them
… if people are interested, start thinking about that
<pchampin> I propose to send a poll today or tomorrow about a weekly slot. The previous poll was explicitly for this week only.
@csarven -- that rules out all the days then :)
<csarven> I meant Monday and Friday, not Monday to Friday =)
ericP: we'll take the timing discussion to the mailing list
<dmitriz> @csarven - and earlier in the call, eric & aaron mentioned that Tue-Weds-Thu are taken up