Re: [Uri-review] [IANA #1270959] Registration of dhttp Schema name (uri-schemes)

[Graham Klyne <[e-mail]>]

Hi all,

(IANA cc'ed, but no action requested at this time.)


I've drafted a possible note for addition to the provisional registration 
template for 'dhttp' [1]

[1] https://www.iana.org/assignments/uri-schemes/prov/dhttp

The intent is not to make a definitive statement about the goodness or badness 
of the scheme, but simply to act as a health warning of potential issues, and to 
highlight that its publication is NOT a recommendation for its use by the IETF.  
Does this look reasonable?

Proposed note follows.

#g


## IANA reviewer note

This scheme has been provisionally registered under the “first come first 
served” policy set out in RFC 7595.  This means that the scheme has not been 
formally reviewed in the IETF, and is not recommended by the IETF for general 
use on the open Internet.

Further, the `dhttp` scheme is considered harmful by some in the IETF [R1] for 
at least the following reasons:

1. it creates confusion around the existing HTTP and HTTPS schemes.
2. ‘dhttp’ is defined to be an alias of the provisionally registered ‘web3’ 
scheme, and URI aliases are, in general, considered harmful to the World Wide 
Web [R2].
3. the activity known as “web3” [R3] has attracted much criticism, and is 
considered by some to be an attempt to insinuate controversial blockchain 
technology into the fabric of the World Wide Web.  Using the name ‘dhttp’ 
appears to be an attempt to legitimise web3 by association with HTTP (the 
current World Wide Web’s primary interaction protocol).

This note is intended to draw attention to potential problems that might arise 
if this scheme were to be used widely on the open Internet, and does not itself 
constitute a position by the IETF or IANA on the considerations raised.

[R1] https://mailarchive.ietf.org/arch/msg/uri-review/4Sj8k6rLZzqZsGgMEe6fLM-wE4U/

[R2] https://www.w3.org/TR/webarch/#uri-aliases

[R3] https://en.wikipedia.org/wiki/Web3

]]



On 20/04/2023 11:34, Graham Klyne wrote:
>
> I see two possible ways forward:
>
> 1. There is an "escape hatch" clause in the registration procedure that allows 
> the IESG to be final arbiter of any contentious registration.
>
> 2. As scheme reviewer, I can request a note be added to the registry entry 
> pointing out that the scheme is contentious, the reasons why, and as such is 
> NOT RECOMMENDED for use on the open Internet.  I would be reluctant to do so 
> on my opinion alone, but I'm seeing sufficient concern expressed here for that 
> to be a reasonable request.
>
> Personally, I think the latter is preferable, for reasons that Ted mentions in 
> a later email.  There are a number of provisionally registered schemes that 
> got snuck in un-noticed before we set up the process of sending notifications 
> of provisional registrations to this list (following the last London  IETF 
> meeting), and I'd be inclined to request a similar note be added to the 'web3' 
> scheme.
>
> #g
>
>
> On 19/04/2023 10:18, Ted Hardie wrote:
>> Hi Roy,
>>
>> The current list of requirements for provisionals is in RFC 7595, Section 4:
>>
>> The scheme name must meet the syntactic requirements ofSection 3.8.
>>
>>    o  There must not already be an entry with the same scheme name.  In
>>       the unfortunate case that there are multiple, different uses of
>>       the same scheme name, the Designated Expert can approve a request
>>       to modify an existing entry to note the separate use.
>>
>>    o  Contact information identifying the person supplying the
>>       registration must be included.  Previously unregistered schemes
>>       discovered in use can be registered by third parties (even if not
>>       on behalf of those who created the scheme).  In this case, both
>>       the registering party and the scheme creator SHOULD be identified.
>>
>>    o  If no permanent, citable specification for the scheme definition
>>       is included, credible reasons for not providing it SHOULD be
>>       given.
>>
>>    o  The scheme definition SHOULD include clear security considerations
>>       (Section 3.7) or explain why a full security analysis is not
>>       available (e.g., in a third-party scheme registration).
>>
>>    o  If the scheme definition does not meet the guidelines laid out in
>>       Section 3, the differences and reasons SHOULD be noted.
>>
>> While it may be the case that using 'dhttp' implies something to humans about 
>> the relationship toother schemes, it meets the current test that "there must 
>> not already be an entry with the same scheme name".  As you will no doubt 
>> recall, we loosened the registration of provisionals in this way because 
>> folks were minting URI schemes without registration and the risk of collision 
>> was getting worse as a result.
>>
>> I am not as clear, though, about whether this registration is intended to 
>> deprecate web3 (which is also a provisionally registered URI scheme) so that 
>> web3 could be marked historic.  If that is the case, we could at least 
>> eliminate the alias scheme issue which you note below.
>>
>> Just my personal opinion, of course,
>>
>> Ted
>>
>> On Tue, Apr 18, 2023 at 4:45 PM Roy T. Fielding <[e-mail]> wrote:
>> >
>> > Is there a way that we can block provisional registrations that are 
>> actively harmful?
>> >
>> >   1) this is abusing the existing http and https schemes;
>> >   2) alias schemes are harmful, in general; and,
>> >   3) web3 is a scam that we shouldn't make respectable by association with 
>> HTTP.
>> >
>> > .....Roy
>> >
>> >
>> > > On Apr 17, 2023, at 4:14 PM, Sabrina Tanamal via RT 
>> <[e-mail]> wrote:
>> > >
>> > > Hi Qi,
>> > >
>> > > We've added provisional URI scheme dhttp to the registry:
>> > >
>> > > https://www.iana.org/assignments/uri-schemes/prov/dhttp
>> > >
>> > > Registry: https://www.iana.org/assignments/uri-schemes
>> > >
>> > > Per the designated expert for URI Schemes registry, we're also notifying 
>> the [e-mail] mailing list upon completing a provisional registration.
>> > >
>> > > Best regards,
>> > >
>> > > Sabrina Tanamal
>> > > Lead IANA Services Specialist
>> > >
>> > > On Mon Apr 17 03:05:35 2023, [e-mail] wrote:
>> > >> Hi Amanda,
>> > >>
>> > >> We would like to register dhttp:// schema with the following information
>> > >>
>> > >> Schema name: dhttp
>> > >>
>> > >> Status: Provisional
>> > >>
>> > >> Applications/protocols that use this scheme:
>> > >>
>> > >> This schema dhttp:// is the alias of schema web3://
>> > >>
>> > >> Contact:
>> > >>
>> > >> Qi Zhou
>> > >> 55 E 3rd Ave, San Mateo, CA 94401
>> > >> mailto: [e-mail]
>> > >>
>> > >> Change controller:
>> > >>
>> > >> Qi Zhou
>> > >> 55 E 3rd Ave, San Mateo, CA 94401
>> > >> mailto: [e-mail]
>> > >>
>> > >> References:
>> > >>
>> > >> A draft specification can be found at
>> > >> https://eips.ethereum.org/EIPS/eip-4804 (replacing web3:// with dhttp://)
>> > >>
>> > >> Scheme syntax:
>> > >>
>> > >> "dhttp://" [userinfo "@"] contractName [":" chainid] path ["?" query]
>> > >>
>> > >> Thanks!
>> > >>
>> > >> Best regards,
>> > >>
>> > >> - Qi
>> > >
>> > > _______________________________________________
>> > > Uri-review mailing list
>> > > [e-mail]
>> > > https://www.ietf.org/mailman/listinfo/uri-review
>> > >
>> >
>> > _______________________________________________
>> > Uri-review mailing list
>> > [e-mail]
>> > https://www.ietf.org/mailman/listinfo/uri-review
>>
>> _______________________________________________
>> Uri-review mailing list
>> [e-mail]
>> https://www.ietf.org/mailman/listinfo/uri-review
> -- 
> Graham Klyne
> mailto:[e-mail]
> http://www.ninebynine.org
> Mastodon: @[e-mail]
> GitHub/Skype: @gklyne
>
> _______________________________________________
> Uri-review mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/uri-review

-- 
Graham Klyne
mailto:[e-mail]
http://www.ninebynine.org
Mastodon: @[e-mail]
GitHub/Skype: @gklyne