Re: DNS -> DIDs -- why not VC embedded in HTML?

I personally believe we (the CCG) should issue specs standardizing "DID
Cross-References" using all three options being discussed (DNS, WebFinger,
HTTP/HTML).

Is there any good reason for us not to specify a standard way to do each of
those three cross-reference methods? And maybe others if they are also
useful?

On Sat, Aug 11, 2018 at 10:17 PM, Markus Sabadello <markus@danubetech.com>
wrote:

> I think both approaches have advantages and disadvantages..
>
> For DNS->DID yes you need to be able to update your DNS.
> For HTML->DID you need a web server (there may be use cases where you
> have a domain name but no web server, or where the DNS service is
> considered more reliable than the web server).
>
> Another approach that has been suggested is WebFinger->DID.
>
> For WebFinger->DID you need a web server AND a WebFinger service, but
> WebFinger is an already widely used protocol.
> Also, unlike the HTML->DID approach it supports email-like identifiers
> (acct:user@domain.com), in other words multiple DIDs per domain name.
>
> Markus
>
> On 08/12/2018 01:03 AM, Manu Sporny wrote:
> > On 08/11/2018 02:17 PM, Markus Sabadello wrote:
> >> - I believe DNS names should only ever be used for initial discovery
> >> of a public DID
> > Why couldn't you just serve something from the root page of the website.
> > Embed a Verifiable Credential stating that "Domain example.com is run by
> > did:example:1234." ... hundreds of millions of sites are doing this with
> > schema.org today.
> >
> > Just copy-paste that into the HTML... Google and Microsoft search
> > crawlers would just immediately pick up those VCs.
> >
> > That's not to say that you shouldn't pursue the DNS approach, Markus...
> > just that it may be easier for web developers to just dump some JSON-LD
> > in their HTML page than it would be to get their organization to update
> > their DNS records.
> >
> > -- manu
> >
>
>
>
>