Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • 01:08
    gibsonf1 commented #288
  • 01:05
    acoburn commented #288
  • 00:18
    acoburn commented #291
  • Jul 27 19:50
    kjetilk labeled #291
  • Jul 27 19:50
    kjetilk labeled #291
  • Jul 27 19:50
    kjetilk opened #291
  • Jul 27 19:50
    kjetilk labeled #291
  • Jul 27 18:39
    kjetilk commented #220
  • Jul 27 18:36
    kjetilk demilestoned #125
  • Jul 27 18:36
    kjetilk commented #125
  • Jul 27 18:21
    kjetilk commented #125
  • Jul 27 18:16
    kjetilk commented #288
  • Jul 27 18:06
    kjetilk unlabeled #267
  • Jul 27 13:48
    csarven labeled #290
  • Jul 27 13:48
    csarven milestoned #290
  • Jul 27 13:48
    csarven opened #290
  • Jul 27 13:46

    csarven on trust-between-owners-same-origin-multiple-storages

    Trust between owners on same or… (compare)

  • Jul 27 13:39
    kjetilk assigned #267
  • Jul 27 13:32
    timbl commented #267
  • Jul 27 13:29
    kjetilk commented #267
Jeff Zucker
@jeff-zucker
No, but the specifications mostly deal with the semantics and leave it as much as possible up to the implementation to decide the details. This is beneficial because it means implementers can innovate and are not bound by lots of rules and are not restricted to file systems for storage, etc.
Daniel B.
@DanielBakas
Agreed. This is very cool for developers. I just hope in a few years users are not discouraged from using Solid because their root directories are filled with folders they don't recognize haha
Maybe the same innovation and creativity could be achieved only by moving everything we develop to "app/data/{application name}" folders
It would still be the same but we would be doing something nice for the user if they ever want a clean root directory to work with
Food for thought 😋
Sarven Capadisli
@csarven
Jeff Zucker
@jeff-zucker
@csarven 404 on the link to sp/TR
Mark Foster
@mikcaweb
Where will the discussions be around picking one of the four versions of RDF (RDF v1.0, RDF v1.1, RDF*, and N3) for the Solid Specification?
Jeff Zucker
@jeff-zucker
@csarven the only solidproject.org link in #284 that isn't 404 is the persistence policy :-)
Sarven Capadisli
@csarven
@jeff-zucker Publication follows merging.
Jeff Zucker
@jeff-zucker
well, that makes sense :-)
sorry
congrats!
Aaron Coburn
@acoburn
@DanielBakas when you have a structure where a particular name is assicated with, for example, a particular app, you end up with a lot of implementation-specific assumptions baked into your whole model. This conflates application data with particular locations, which is not something the specification should say anything about. One Solid server may use DIDs for identifiers, which explicitly separates the notion of resource identity from resource location; another may enforce entirely opaque (e.g. UUID) identifiers for resources. That and many other structures are all possible with a well-defined semantic structure that is understandable across client applications. That is the level where interoperability will reside, not in having a particular mapping app “reserve” the /map top level directory. Such a structure not only would require a centralized registration/authority, but it also does not scale.
LievenVDP
@LievenVDP
Hi, I got directed here from app-development,
We are currently about to kick off development on an application which will use SOLID heavily. The application will have search and matching functionality and thus we will need to keep certain data cached on our end to perform these actions. I would like a way where i could register a url with a pod provider and the provider would call that url when access gets revoked (so i can clear the stored data we used for searching on our end), sort of like a simple webhook. The alternative is that i would have to test for connections every x time which seems like alot of overhead for such a simple thing. The other thing we would really love to see is a way to trigger pod registration through our application with an api or something where the service would then send the info back to us . We want all our users to use the pods but our fear is that without a seamless integration the use of SOLID would scare away alot of potential users.
We would be interested in helping develop these features Is there animo for these features (to us these seem like basic features most commercial implementations would need) and who would i need to talk to about this further?
Fred Gibson
@gibsonf1
@DanielBakas For the /apps/data idea, what this means is that an app adds that to the privateType registry for the user, which as an example with Media Kraken, would be an entry like this:
<https://frederick.trinpod.us/profile/settings/frederick.privateTypeIndex#3b2fa422-5591-4924-9c2a-002c96653ccf> 
    a solid:TypeRegistration ;
    solid:forClass <https://schema.org/Movie> ;
    solid:instanceContainer <https://frederick.trinpod.us/apps/data/movies/> .
Fred Gibson
@gibsonf1
@DanielBakas So Media Kraken first checks each time the app is used in the user's pod for an entry in the privateTypeIndex for the class schema:Movie, and if its there, "follows its nose" to the solid:instanceContainer , and if that entry is not there, it creates a new one for that user
so that is how you implement the idea of apps starting to use /apps/data/ to locate info
And whats nice with this is that media kraken and other apps that have traditionally used the root, can after checking for a user's location, if not there, have the new user use the new location. So the app continues to work for the original location and the new location depending on that entry in the index
Fred Gibson
@gibsonf1
And then the other piece is finding where the index is, which can be found by looking at the user's webid request:
@base <https://frederick.trinpod.us/i> .
@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
@prefix frederick: <https://frederick.stage.graphmetrix.net/node/> .
@prefix acl: <http://www.w3.org/ns/auth/acl#> .
@prefix vcard: <http://www.w3.org/2006/vcard/ns#> .
@prefix foaf: <http://xmlns.com/foaf/0.1/> .
@prefix ldp: <http://www.w3.org/ns/ldp#> .
@prefix solid: <http://www.w3.org/ns/solid/terms#> .
@prefix schema: <http://schema.org/> .
@prefix space: <http://www.w3.org/ns/pim/space#> .
<https://frederick.trinpod.us/i> 
    acl:trustedApp frederick:t_at , frederick:t_2p ;
    foaf:name \"Fred Gibson\"^^<xsd:string>  ;
    ldp:inbox <https://frederick.trinpod.us/inbox/> ;
    a schema:Person , foaf:Person ;
    solid:account <https://frederick.trinpod.us/> ;
    solid:oidcIssuer \"https://trinpod.us\"^^xsd:string ;
    solid:privateTypeIndex <https://frederick.trinpod.us/profile/settings/frederick.privateTypeIndex> ;
    solid:publicTypeIndex <https://frederick.trinpod.us/profile/settings/frederick.publicTypeIndex> ;
    space:preferencesFile <https://frederick.trinpod.us/profile/settings/frederick.preferences> ;
    space:storage <https://frederick.trinpod.us/> ;
    vcard:hasEmail \"frederick@graphmetrix.com\"^^xsd:string ;
    vcard:url frederick:t_59 .
Daniel B.
@DanielBakas
Okok I think I get it. Would it still be desirable to specify a location for the application's data? I'm all too new to all of this so thanks to everyone who has been generous enough to explain!
Fred Gibson
@gibsonf1
yes, you basically add the location you want in the privatetypeindex
Daniel B.
@DanielBakas
Cool! All that's missing is (if considered relevant and appropriate) to designate a standard folder for settings and all data from our applications that are not relevant to the user (right?)
elf Pavlik
@elf-pavlik
You might like to take a look at new work in https://solid.github.io/data-interoperability-panel/specification/ which also addresses authorizing applications
Daniel B.
@DanielBakas
Thank you!!
elf Pavlik
@elf-pavlik
BTW I'm currently working on TypeScript implementation of that spec, official repo: https://github.com/janeirodigital/sai-js (first release later this month)
Daniel B.
@DanielBakas
Hi guys, question here: if I want to have public data that everyone can access on my application (regardless on wether the users have pods or not) should I be using a pod for an organization or should I use a triplestore like Fuseki? :O
Sarven Capadisli
@csarven
You can use any HTTP server.
Daniel B.
@DanielBakas
But I mean, where should the data be stored? In a pod or a SPARQL endpoint?
elf Pavlik
@elf-pavlik
I don't think it's either-or https://linkeddatafragments.org/concept/ illustrates it very nicely, i think LDP fits as 'subject page'
bblfish
@bblfish:matrix.org
[m]
csarven: you sent me a message last week but Matrix has some problems decrypting the message.
bblfish
@bblfish:matrix.org
[m]
Fred Gibson
@gibsonf1
Would it be possible for the group to assign a unique URI to each Solid spec item to use for linking specs to tests and other materials?
and those spec items can be linked such that a requirement that part of a larger requirement can be contained by it
image.png
Jeff Zucker
@jeff-zucker
@gibsonf1 I believe that the Technical Reports are the authoritative place to cite. Specific items can be found with address fragments e.g. https://solidproject.org/TR/protocol#writing-resources
Not sure about the other parts of your question.
Pete Edwards
@edwardsph
@gibsonf1 This is something that we've been looking at too. It's really important that tests can be traced to the spec requirements they relate to. As @jeff-zucker said, you can point to broad sections in a spec but @csarven has started some work on a vocab we could use to annotate the specs at a more granular level. Then tests can be tied to specific requirement statements.
Fred Gibson
@gibsonf1
That would be excellent, just need a unique URI for each testable requirement of the spec
Fred Gibson
@gibsonf1
Maybe simply add _number to each sub requirement, so that requirement #1 "When a server supports the HTTP PUT, POST and PATCH methods [RFC7231] this specification imposes the following requirements: [Source]" becomes https://solidproject.org/TR/protocol#writing-resources_1
Vincent
@Vinnl
Hey all, for those having trouble keeping up with the forum, there's what looks like a suggestion for the spec here: https://forum.solidproject.org/t/the-client-key-should-not-be-bound-to-the-refresh-token/4516
Kjetil Kjernsmo
@kjetilk
thanks for opening #288, @RubenVerborgh
Fred Gibson
@gibsonf1
I was just looking at the great work at https://github.com/solid/conformance-test-harness and noticed some turtle referencing a spec vocab on w3: https://www.w3.org/ns/spec#Specification . Does anyone know where that whole ontology can be downloaded from?
Pete Edwards
@edwardsph
That is the possible namespace for a vocab that Sarven has started. The source is here: https://github.com/solid/vocab/blob/specification-terms/spec.ttl
Fred Gibson
@gibsonf1
Thanks!
We've been using sio:SIO_000090 for Specification as the sio ontology is the best ontology we've found that tries to model the full interconnected world. http://visualdataweb.de/webvowl/#iri=http://semanticscience.org/ontology/sio.owl
Fred Gibson
@gibsonf1
@csarven The Spec ontology is looking really good