Re: [Multiformats] New versions of Multibase and Multihash published

[Melvin Carvalho <[e-mail]>]

út 5. 9. 2023 v 20:50 odesílatel Manu Sporny <[e-mail]>
napsal:

> On Tue, Sep 5, 2023 at 2:35 PM Melvin Carvalho <[e-mail]>
> wrote:
> > Could you clarify what you're referencing in regards to Bengo and the
> associated controversy?
>
> Yes, Bengo said this:
>
> On Tue, Sep 5, 2023 at 2:15 PM [e-mail] <[e-mail]> wrote:
> > Melvin, I believe both dweb: and ipfs: URI schemes are registered with
> IANA and use CIDs.
>
> to which you responded with this:
>
> On Tue, Sep 5, 2023 at 2:17 PM Melvin Carvalho <[e-mail]>
> wrote:
> > there's s degree of controversy around dweb: and I'd rather not use that
>

The "dweb:" scheme might not suit our needs; let's discuss suitable
alternatives. I believe that's a separate issue. Please advise if otherwise.

"ipfs" serves as a transport mechanism via a DHT. I'd prefer an identifier
independent of IPFS for stability, especially since "ipfs:" has evolved
over time.


>
> > I'm ready to utilize your work and register a provisional cid: scheme or
> urn:cid: scheme, based on your preference.  Should be a simple enough copy
> and paste.
>
> I suggest you start with a spec that defines the "cid:" scheme (or
> re-use the dweb: scheme) if what you're looking for is a URL scheme to
> express content identifiers.
>

It looks like the "cid:" scheme is already taken. ci: however is not.

I initiated discussions on associating a URI with multihash on the W3C
list, leading to a consensus for "mh:" over "multihash:". Using a scheme
that isn't reflective of wide deployment and lacks backward compatibility
seems counterproductive. Would "urn:cid" be suitable? "Multihash:" might
cause naming issues. I'm open to suggestions, but it should be agreeable to
your group.


>
> > Given the widespread deployment of CID strings, this approach may not be
> optimal. It seems to divide standardization and might not be backward
> compatible.
>
> Hrm, how so?
>
> > However, if you feel strongly about this direction, I'm open to
> considering two standardization paths.
>
> Yes, CIDs are out of scope at present and I'm afraid that if we put
> them in scope at this late juncture that we'll get objections.
> However, a separate standardization path that defines a new CID URL
> scheme, taken through dispatch, and then covered by an existing group
> (or added to the charter for this group after our first iteration)
> sounds like a viable path forward.
>
> -- manu
>
> --
> Manu Sporny - https://www.linkedin.com/in/manusporny/
> Founder/CEO - Digital Bazaar, Inc.
> https://www.digitalbazaar.com/
>