Re: [urn] DOI URN namespace registration

[John R Levine <[e-mail]>]

On Sun, 24 May 2020, Dale R. Worley wrote:
> "John Levine" <[e-mail]> writes:
>> ...
>> Given that DOIs are usually resolved by looking up URLs, it is clear
>> that in practice they are compatible with URLs because they *are*
>> URLs.
>
> Unless they've retroactively removed non-ASCII characters from being
> valid in DOIs, then it's certain that DOIs *aren't* URLs, and quite
> possible that nobody has ever properly specified how DOIs are mapped
> into URLs for lookup.

That's possible but on the other hand, people have been looking up
DOIs as URLs forever so there is clearly a convention that works.
While the handle spec says that the local part can be any UTF-8, I
can't ever recall seeing a DOI that wasn't ASCII. Given that DOIs are
opaque identifiers, there's no obvious benefit to using non-ASCII.

>> Beyond the case sensitivity thing which I think is a red herring, I
>> would ask why they're asking for a DOI URN rather than a handle URN
>> since they're the same namespace.
>
> It certainly seems preferable to solve the general case rather than just
> the special case.  But I notice that though I can't remember anything
> about "the handle system", my subconscious supplies a frisson when I
> read that phrase.

There is certainly a lot of unpleasant political history.  The spec
is in RFCs 3650 and 3651.  The first component of a handle is the authority,
and 10.* authorities are DOIs.

I don't think it's unreasoanble to have a URN namespae for handles or
DOIs, and I don't think it should be hard to make it compatible with
existing DOI practice, but I do think there's loose ends to be tied
down

Regards,
John Levine, [e-mail], Taughannock Networks, Trumansburg NY
Please consider the environment before reading this e-mail. https://jl.ly