Re: W3C Solid Community Group

On Thu, 21 Mar 2019 at 11:42, Michiel de Jong <michiel@unhosted.org> wrote:

> Great, see you all in an hour and a few minutes from now! :)
>
> I realize the agenda for this one is now full, but since I had already
> typed this email, I'll just ask if we can add three topics to the end of
> the queue, and what we don't have time for can flow over to the next
> meeting:
>
> * cross-server testsuite: we agree that we need something like
> https://github.com/w3c/ldp-testsuite +
> https://github.com/csarven/ldn-tests + WAC tests. Who wants to work on
> this?
>

+1 on a test suite

We are in a sort of node solid server bubble at the moment, which is not
healthy.

node / javascript is one of those things that is quite polarizing.  At one
time most work went into our php server, at one time go, and most recently
node and that has captured the mind share of the current development team

There is a real danger that the node implementation becomes labeled
reference and the spec is taken from that.  This is a mistake.  For example
the OIDC work was commissioned to be an extra addition to live along side
existing auth and provide a point of flexibility.  What happened in
practice is it became the fox in the hen house and tried to replace
existing methods, which for example work perfectly fine on server and
command line, and we lost webid delegation in the process.

Result is that we lost interop and all our apps broke.  We are now
scrambling to fix this and have been for a year.  Just one of the many
issues.

So, yes!  This is really important!  Thanks for bringing it up.


>
> * data collections: a collection (term coined by Justin yesterday) would
> be something like "my contacts" or "my family photos". It's not necessarily
> in  one LDP container, and  it's not necessarily one RDF type. It could
> also be data from a specific source,  or data you use in a specific way,
> with specific people, or for a specific purpose.  We've been using the type
> index for data  discovery, but this is also related to access control. We
> need to think more about this concept, since it's key to the sharing UX (
> as in "do you allow this app to acess your ... colllection?").
>

Nice idea.


>
> * trusted apps UX: partially this is just the technical implementation
> details of the previous topic ("given that the user wants  to  give app X
> access to collection Y, how does that work step-by-step with webid and
> wac?). This is not specified in detail in the spec but basically the
> implementation we have now in NSS needs some work, and we probably want to
> standardize this flow a bit across implementations
> https://github.com/solid/node-solid-server/issues/1142. Also
> https://github.com/solid/solid-spec/issues/142 which came out of the last
> meeting.
>

Yes, very important too.  Also outstanding question is whether or not, or
when, we should turn on trusted apps.  Is it needed just yet?


>
> My 2ct,
> Michiel
>
>
> On Thu, Mar 21, 2019 at 6:39 AM Mitzi László <mitzil@inrupt.com> wrote:
>
>> Hi W3C Solid Community Group,
>>
>> There have been three conversations over the email in the past couple of
>> weeks:
>>
>>    1. Ontologies (Timo)
>>    2. Identity Model (Timo)
>>    3. Header to allow WebID TIS on servers (Melvin)
>>
>> If you would like to discuss them on the W3C Solid Community Group call
>> please let me know. The agenda for the upcoming meeting is set, if we have
>> time and you feel prepared we could dip into the additional conversations
>> today. What would be helpful would be to define which parts of the spec are
>> relevant to this conversation and where in the spec do you think additional
>> work would be beneficial.?
>>
>> Melvin - in réponse to your question earlier, the purpose of the document
>> on The Default was to demonstrate what I have in mind in preparation for an
>> official conversation with our group in the upcoming call to decide how to
>> move forward together.
>>
>> Mitzi
>>
>>

Received on Thursday, 21 March 2019 10:55:31 UTC