solid/specification

Repository: https://github.com/solid/specification - Technical Reports: https://solidproject.org/TR/ - Code of Conduct: https://github.com/solid/specification#code-of-conduct

Sarven Capadisli

@csarven

Good news everyone!

We're (again) in contact with the W3C Web Platform Incubator CG and this time we're looking into exchanging our notes/work over a meeting:

WICG/WebID#54

(
I've mentioned WICG's work/initiative around "WebID" in the authentication-panel awhile back:
https://gitter.im/solid/authentication-panel?at=5f96a28d3d172d78b39bf4bd
https://gitter.im/solid/authentication-panel?at=5f9aebabd5a5a635f28b8415
)

Date:
February 22 (Monday) at 17:00 GMT (for a ~2h session, TBD) is the proposed date for the meeting.

Sessions:

There is no particular format (yet) other than to get an understanding of each others' work. So, brief presentations with a lot of Q&A is a possibility.
Anyone is welcome to attend. Please join at least one of the CG's if not already.
Sessions may/will be minuted.

Todos:

Confirm whether you can attend. Authors/editors of specs, especially WebID, SOlid-OIDC, people working on implementations.. please chime in.
Solid CG attendees to read https://lists.w3.org/Archives/Public/www-archive/2020Oct/att-0010/The_Web_Platform__Privacy_and_Federation_-_TPAC.pdf and consider having some questions prepared.
..

Justin Bingham

@justinwb

This is great @csarven - confirmed I can be there :white_check_mark:

Sarven Capadisli

@csarven

:) Exciting, isn't it?

Justin Bingham

@justinwb

it is!

Henry Story

@bblfish

Have added it to my calendar. Will need to catch up on what they are doing too now!

Sarven Capadisli

@csarven

Will need to catch up on what they are doing too now!

Story of my life.

Aaron Coburn

@acoburn

@csarven I can attend.

Emmet

@emmettownsend

@emmettownsend
While moving to the next phase of our development on Solid products we found we needed a flexible, robust test suite so that we can test for conformance against the specificaiton. That led me to write a blog with the help of @edwardsph and @RubenVerborgh . Happy to get feedback if you get a chance to have a read. Thanks.
https://inrupt.com/blog/conformance-test-suite

Jordan Shurmer

@JordanShurmer

yes! This is awesome @emmettownsend

Fred Gibson

@gibsonf1

Nice post @emmettownsend , and I definitely vote RDF wherever possible

Justin Bingham

@justinwb

:clap: :clap:

Jordan Shurmer

@JordanShurmer

I would definitely use such a suite of tests btw.
I have a Rust implementation I'm starting to get back to..
wrote my own tiny tests so far, but something like that would be very useful

Justin Bingham

@justinwb

really like the structure of this framework @emmettownsend - this will also do nicely for the shape tree conformance testing that’s about to get underway

Sarven Capadisli

@csarven

Issue/discussion from 2019-07 on supporting RDFa in specifications (based on prior experience): solid/specification#6

One of the reasons why bikeshed/respec (compiling) pipeline is not used for /TR/ , /TR/ecosystem , /TR/protocol , and soon /TR/web-access-control , (and hopefully other reports will follow along) is that it doesn't fit the Solid pipeline ie. what Solid servers (eg. NSS, CSS, ESS..) and applications can do (eg. dokieli). The template is already in place.. and after some more restructuring, we'll have fine-grained statements (eg. all the bits of a requirement). The LDN spec did it up to a point of identifying and describing the requirements - I ran out of time back then to do more with the test suite. Now we have that possibility.

See also Linked Specifications, Test Suites, and Implementation Reports: https://csarven.ca/linked-specifications-reports (2017) for an overview on existing work and how all major components fit together.

Henry Story

@bblfish

For our march meeting with the Credentials community I think I have found a very nice way for us to interact with them via @dmitrizagidulin's did:web proposal. I wrote up details on https://github.com/solid/specification/issues/217#issuecomment-777375570

Sarven Capadisli

@csarven

OK, so, I didn't hear back from anyone re topics with CCG.. besides @bblfish's solid/authentication-panel#126 ... so going to stick to their suggestions, bblfish's, and mention zcap perhaps.

As for meeting with WICG, I take it that we'll stick to the suggested date (Feb 22) - will keep everyone posted if there is any specific agenda beyond intro to Solid's use of WebID, Solid-OIDC, and WICG's WebID.

Henry Story

@bblfish

I am developing some interesting ideas starting from @dmitrizagidulin ideas on did and solid that he posted here https://github.com/solid/specification/issues/217#issuecomment-777101431 nearly a couple of years ago to talk to the CCG about

Sarven Capadisli

@csarven

Right. I think that'll be covered naturally.. (IMO) but will mention it.

Sarven Capadisli

@csarven

@/all CCG announcement for the meetings with SolidCG: https://lists.w3.org/Archives/Public/public-credentials/2021Feb/0062.html

Fred Gibson

@gibsonf1

Does anyone know if there is official posix or other ontology for a file name and folder. We found this https://www.w3.org/ns/posix/stat, but knowing what standard predicate is being used for a file name would be very helpful

Sarven Capadisli

@csarven

BTW, the ns is in http, not https

Sarven Capadisli

@csarven

Seen https://lov.linkeddata.es/dataset/lov/terms?q=filename&type=property ? Probably best to decide on what it is supposed to be a property of. For example, is it to say this resource is mapped to a /path/to/file on a local system?

Fred Gibson

@gibsonf1

In our case we have a RDF node that represents a file, and there is a version property that points to a file. We use a tag property to represent the file name and use that tag, and the tags of containers, to build a path based uri to map back to that node. The issue is that a file and container name need url encoding to accept spaces etc, so we want to use a special version of the tag that indicates urlencoding and just wondering if there is some property typically used for the file name (and container) already in use

Fred Gibson

@gibsonf1

We now need to add search to our App ideally using an auto-fill approach where as you type each letter the search is refined. Ideally this would be hitting a sparql query that has results filtered based on permissions, but is there a provision in Solid for doing sparql queries to find information or should we just do a separate solid secured sparql endpoint for that?

Aaron Coburn

@acoburn

@gibsonf1 keep in mind that not every Solid server uses files or even local filesystems

Aaron Coburn

@acoburn

Including a SPARQL query endpoint is an implementation decision, but it is not required by the current protocol spec document

Fred Gibson

@gibsonf1

regardless of implementation, it would be good to have some kind of search endpoint that servers could respond to

Aaron Coburn

@acoburn

There have been various conversations over time, but this captures a lot of the current thinking: https://github.com/solid/specification/issues/162#issuecomment-608632015

Fred Gibson

@gibsonf1

https://github.com/solid/specification/issues/162#issuecomment-778699371

Fred Gibson

@gibsonf1

Link above = How about a very simple standard endpoint on solid in the format of /search that takes a string of characters, similar to typing in a google search box, that the Solid Server can respond to in whatever form the client requests such as turtle etc?

Fred Gibson

@gibsonf1

For retrieving ontology for use in our app, we have an endpoint on the Server /ns that takes requests in the form of /ns/foaf/Person ( ns/prefix/term ) to then resolve on server side to the uri and return ontology triples from that request. The issue of course is that with ontology, you can't use those uri's directly to retrieve any custom ontology built around the original or in many cases, the uri's don't resolve at all. Does Solid have any plans for the ontology request issue?

Sarven Capadisli

@csarven

? Why can't applications GET the vocabs/ontologies from source? What's the use case?

Henry Story

@bblfish

yes, ontologies should be available via GET if well designed. But PLEASE build caching systems

Emmet

@emmettownsend

I think it will be necessarily to have a few standard endpoints. However rather than hard code them we should think about using something like a .well-known/solid endpoint where a client can discover the endpoints supported by a server.

Henry Story

@bblfish

The whole notion of "endpoint" smaks somehow of SOAP, so I tend to think one should avoid it.

Emmet

@emmettownsend

for example . a registration endpoint.

Henry Story

@bblfish

registration for what?

Emmet

@emmettownsend

allows the creation of pods without requiring a user and a browser. and makes it standard across servers.

Search is anothers example

certainly not anythng like SOAP though

Henry Story

@bblfish

That sounds a bit like an index...

Emmet

@emmettownsend

index is just an implementation detail

its really about discovery

how do I create a pod on this service.

how do I search on this service

Henry Story

@bblfish

I think containers and resources should provide their own query link...

Emmet

@emmettownsend

ok so providing a link header to allow a client discovver how it queries a container is certainly one way. But we also need to think about cross pod searches and corss ecosystem searches

similarly one could have a link header from a server to discover the pod resistration end point.

Some of these things will not change and there may be push back around adding lots of link headers. Hence another approach would be a .well-known/solid resource

Where communities thrive

People

Repo info

Activity