[urn] Re: Additional feedback re: the glue URI scheme

[[e-mail]]

Pamela Dingle <[e-mail]> writes:
> There is a company called Weinerberger AT. They are one of Europe's
> largest manufacturers of roof tiles and bricks.  At least, that is
> what the company is called now.  It was first registered in 1819, in a
> commercial register called the Wien Handelsregister (Vienna Commercial
> Register).  That register was obviously not digital, but there is an
> entry in that register for "Wienerberger Ziegelfabriks- und
> Baugesellschaft" (Vienna Brickworks and Construction Company).

Thanks for providing this example, as I think it will help clarify
things.

In this case, the type of identifier is "Vienna Commercial Register", or
some incarnation of that register.  If anyone starts caring about it as
a means of identification, I expect that the Register will be given a
unique identifier within this space.

But it seems to me that the *identifier* issued by the Register is
likely to be (to have been!) the text string "Wienerberger
Ziegelfabriks- und Baugesellschaft".  Looking at the ChatGPT summary of
"issuerAlsoKnownAs", it looks complicated enough that I won't attempt to
analyze it now.  But given that we have a business whose name is
recorded into a historical register, how can we encode that information
into an issuerAlsoKnownAs?  The only datum in the register is the name
and the definition of issuerAlsoKnownAs seems to require that the "id"
is a globally unique URI.

Supporting this gets rather messy, as even the gimmick I proposed --
considering the name as a string and turning it into a "data:" URI --
doesn't suffice because there's no guarantee that the name is unique
across all historical registries.

Dale