Re: W3C Solid Community Group Call in Details

On Thu, 21 Mar 2019 at 14:36, Melvin Carvalho <melvincarvalho@gmail.com>
wrote:

>
>
> On Thu, 21 Mar 2019 at 14:08, Michiel de Jong <michiel@unhosted.org>
> wrote:
>
>> The minutes are here: https://www.w3.org/2019/03/21-solid-minutes.html
>>
>
> Good minutes, thanks!
>
> Dmitri, I've also been discussing a functional did spec with manu
>
> There are some important design decisions which give different
> functionality, and the devil is in the details.
>
> It strikes me that there is a LOT of devil in the detail, which will
> produce very different types of eco system.
>
> The good news is that the DID specs lend itself to be able to work on more
> than one, but it might make more sense to try and come up with a common
> design goals
>
> Examples being around the type of key material used, the serialization of
> that key material, conversion via a trap do function (aka hash) to an UUID,
> one-one relations or one-many relations, whether to use a block chain /
> distributed ledger / or LDPC, bootstrapping a central registry, discovery,
> importantly for solid turtle vs json-ld support and how that relates to
> relative URIs and so on.
>

"trap do" should read "trap door", also I should mention life cycle
aspects, which are a whole spec in itself -- ie what happens when you lose
your private key ... this is a 10 year old problem in solid, and may not be
solved quickly, but should not hold back making some useful specs ...


>
> What I suggest is we start with something very simple which bootstraps our
> existing infrastructure.  For example having a content addressable document
> (did) for each key in a webid on a one-to-one (isomorphic) basis seems to
> be low hanging fruit, then allowing the CRUD operations to happen via the
> HTTP verbs and PATCH with our own dialect of sparql update.  I suppose that
> will mean having to write up sparql-update-solid someplace.
>
> With that part in place, we could perhaps look at various other directions
> we might like to take it, as a more formal spec.  But having a long lasting
> document for public keys is in itself valuable, keybase are an example who
> found unexpected reuse in this respect.
>
> Thoughts?
>
>
>>
>> On Thu, Mar 21, 2019 at 8:08 AM Mitzi László <mitzil@inrupt.com> wrote:
>>
>>>  https://zoom.us/j/121552099
>>>
>>