Re: WebID default serialization for WebID 2.x

On Wed, 26 Jan 2022, 02:29 Jonas Smedegaard, <jonas@jones.dk> wrote:

> Quoting Nathan Rixham (2022-01-26 03:09:03)
> > On Tue, Jan 25, 2022 at 11:59 PM Kingsley Idehen <kidehen@openlinksw.com
> >
> > wrote:
> >
> > > On 1/25/22 6:08 PM, Nathan Rixham wrote:
> > >
> > > On Tue, Jan 25, 2022 at 10:58 PM Kingsley Idehen <
> kidehen@openlinksw.com>
> > > wrote:
> > >
> > >> On 1/25/22 4:29 PM, Nathan Rixham wrote:
> > >>
> > >> On Tue, Jan 25, 2022 at 9:23 PM Kingsley Idehen <
> kidehen@openlinksw.com>
> > >> wrote:
> > >>
> > >>> On 1/25/22 1:28 PM, Nathan Rixham wrote:
> > >>>
> > >>> Would a fair definition of a valid WebId then be something like: A
> URI
> > >>> is a valid WebIdentifier if it dereferences to a valid WebId-Profile
> > >>> describing the URI with the minimum set of required properties
> (type, name,
> > >>> public_keys)?
> > >>>
> > >>> A WebID is a resolvable identifier that denotes an agent. It
> resolves to
> > >>> a WebID Profile Document.
> > >>>
> > >>  How do you know it's a WebID before you resolve it?
> > >>
> > >> Good question!
> > >>
> > >> By deciding the want to denote yourself using a given identifier,
> > >> relative to your profile document.
> > >>
> > > Let me rephrase, and suggest looking at it the other way around: given
> a
> > > random IRI <y> how do I know <y> is a webid / refers to an Agent,
> without
> > > first resolving it? and as a sub point, is a "valid" webid?
> > >
> > > You know it is a WebID because the spec says its is a URI that denotes
> an
> > > Agent. That's it, really :)
> > >
> > > "
> > > WebID A WebID is a URI with an HTTP or HTTPS scheme which denotes an
> > > Agent (Person, Organization, Group, Device, etc.). For WebIDs with
> fragment
> > > identifiers (e.g. #me), the URI without the fragment denotes the
> Profile
> > > Document.
> > >
> > > "
> > >
> > I feel something is getting lost in translation.
> >
> > There's no way to know this without doing something first, to implement
> > anything in code it would need to be the opposite way around, "A URI is a
> > WebID if...", there's no way for a system to know it's a WebID without
> > further information, and the spec should define what that information is,
> > and how to get it, the minimum needed to determine that a previously
> > unknown uri <y> as a WebID via some mechanism.
> >
> > "A WebID is a URI with.." doesn't actually help or provide any guidance
> or
> > useful definition, indeed it's entirely impossible to get any further
> > unless you know beforehand that `<y> an :Agent`, but you don't know this
> > until after you've tried to resolve <y>  ..
> >
> > Am I making sense here?
>
> How do you know that https://example.org/ is a web address?
>
> When you throw it at a web browser you test that it *works* as a web
> address, not if it *is* a web address.  If it fails to work then you
> have only really tested the web service, not the address.
>
> You can check if I live at "Foobar street 45", but you cannot test if
> "Foobar street 45" *is* a (real or imaginary) street name.
>
> Does that answer your question - or do you feel I am sidestepping and
> missing your point?


Really simple stuff, WebID spec should define how to differentiate a WebID
from a random URI, otherwise what's the point, it doesn't actually specify
it.

A URI/IRI is a WebID if..
You can determine that a URI/IRI is a WebID by ...

If it doesn't define what it is, and how to determine it is that, what is
the spec?

>