Re: [Uri-review] Wi-Fi Alliance permanent scheme registration -review request

[Graham Klyne <[e-mail]>]

Hi,

Larry's comment raised a flag for me, so I took a look at the referenced I-D [1] 
(per https://mailarchive.ietf.org/arch/msg/uri-review/1NP3Q5xB6JZt6nNxFLa3g7MAnGs).

[1] 
https://datatracker.ietf.org/doc/draft-op3ft-leaptofrogans-uri-scheme/?include_text=1

A touchstone I have in situations like this is to ask the question "what do 
instances of this URI scheme identify?", and in this case I think it's clear 
enough that they identify a "frogans site" (whatever that may be).  So far, so good.


However, reading the draft, I have other concerns:


1. Section 1: it's not at all clear what FrogAns can achieve that is not 
achievable using WWW and appropriate media types.  (Which somewhat supports 
Larry's point.)  There is plenty of "visual rather than text-based 
communication" that can be and is implemented using WWW.  I don't see this as a 
blocker, but it maybe suggests the documentation could be more helpful.


2. Section 2, according to the registration template, is the goto reference for 
scheme semantics.  But looking at this, I'm not seeing much description of 
semantics.

I don't know what the first paragraph of section 2 is actually intended to mean. 
I can't recognize the characterization of URI usage offered here.

The second paragraph talks about referencing the Web from frogans sites, which 
doesn't seem to need a new URI scheme.

The third paragraph talks about referencing frogans sites from the Web, which I 
assume is the intended use of this scheme.  The goals suggested here appear to 
assume very broad deployment of the frogans scheme implementation, which is not 
something I think is likely to happen (especially since the technology, or at 
least its name, appears to be quite proprietary).  For ease of deployment (at 
scale), I think Larry's point about using a MIME type is particularly relevant. 
  Also, RFC7595 points out the potential high cost of deploying new URI schemes [2].

[2] https://tools.ietf.org/html/rfc7595#section-3.1

(I personally think that schemes like "mailto:" set a poor precedent for URI 
scheme use, as they obscure the role of URIs as identifiers and locators of 
resources in the Web.  But that horse being long bolted doesn't mean we can't 
try to do better.)


frogans is described as having a "slide description language" format (FSDL). 
(FSDL even appears to include a format for sending requests to servers.)  Is 
there any reason that FSDL cannot be accessed over the web using HTTP, and 
interpreted/dispatched to a frogans player?  This doesn't mean that HTTP has be 
be used by frogans native applications, but IMO would allow frogans to integrate 
more easily into non-frogans environments by exploiting existing protocols. 
E.g., it could allow things like sending FSDL in email and having the recipient 
client dispatch to a frogans player.

It may well be that there is a good case for having a new URI scheme for 
frogans, but I feel the draft doesn't really explain it.  Understanding why 
HTTP+media type doesn't work just as well for the intended use would help me to 
understand what that case could be.

#g
--


On 06/10/2018 03:30, Larry Masinter wrote:
> I think they might want to reconsider whether they need a URI scheme at all, rather than a MIME type.
> When you have a complex but extensible data structure, it’s more resilient. And if you have
> Situations where you really need to serialize as a URI, you can always do it with a data: URI.
>
> Larry
> --
> http://LarryMasinter.net
>
> From: Gaurav Jain
> Sent: Friday, October 5, 2018 3:24 PM
> To: Ted Hardie
> Cc: [e-mail]
> Subject: Re: [Uri-review] Wi-Fi Alliance permanent scheme registration -review request
>
> Thank you Ted for your helpful suggestions. Wi-Fi Alliance discussed this feedback and think the second of your suggestions on how to have an extensible Bootstrapping Information URI is preferable. We would therefore like to additionally have a separate registry of tokens, initially populated by C, M, I, and K with definitions taken from the description we have provided. As a condition of adding to that registry may we suggest 'Expert Review' as a registration procedure for new tokens with myself, Gaurav Jain <[e-mail]> as the designated expert. Alternate registration procedures are possible if 'Expert Review' is not appropriate in your estimation.
>
> Thanks and regards,
>
> Gaurav Jain
> Senior Manager, Program Technology
> [phone] Office
> [phone] Mobile
>
> Wi-Fi Alliance
> www.wi-fi.org
> http://twitter.com/wifialliance
> www.facebook.com/wificertified
>
> Visit our blog, The Beacon, for Wi-Fi industry topics and trends
>
> From: Ted Hardie <[e-mail]>
> Sent: Tuesday, September 18, 2018 7:45 AM
> To: Gaurav Jain <[e-mail]>
> Cc: [e-mail]
> Subject: Re: [Uri-review] Wi-Fi Alliance permanent scheme registration - review request
>
> Hi Guarav,
>
> Thanks for your message. Poking at the spec, the syntax of the dpp scheme appears to be given as:
> "
> dpp-qr = “DPP:” [channel-list “;”] [mac “;”] [information “;”] public-key “;;”
> pkex-bootstrap-info = [information]
> channel-list = “C:” class-and-channels *(“,” class-and-channels)
> class-and-channels = class “/” channel *(“,” channel)
> class = 1*3DIGIT
> channel = 1*3DIGIT
> mac = “M:” 6hex-octet ; MAC address
> hex-octet = 2HEXDIG
> information = “I:” *(%x20-3A / %x3C-7E) ; semicolon not allowed
> public-key = “K:” *PKCHAR ; DER of ASN.1 SubjectPublicKeyInfo encoded in “base64” as per [14]
> PKCHAR = ALPHA / DIGIT / %x2b / %x2f / %x3d
>
> The channel-list ABNF rule allows a list of IEEE 802.11 global operating class and channel (Annex E of [2]) value pairs to
> be specified. The MAC ABNF rule expresses the MAC address as a string of six hex-octets. The information ABNF rule
> allows freeform information to accompany the public key.
> The bootstrapping information may be extended in future updates of the technical specification. Devices parsing this
> information should ignore unknown semicolon separated components in the dpp-qr and pkex-bootstrap-info instantiations
> to be forward compatible with such extensions."
>
> The last paragraph indicates that this is subject to later extension, but the syntax does not appear to be versioned and there does not appear to be any requirement for the placement of future extensions.  My experience is that this could well result in later interoperability problems.
>
> There seem to be a couple of ways to make that work, if you do not anticipate registering dpp2 and so on.  One would be to include later extensions within the current free-form Information field, declaring a new delimiter,which would be understood by the later version.  Reserving the delimiter now would be useful, if that is your choice.  Another would be to generalize this slightly so that the syntax of the scheme is simply (scheme name) ":" followed by a series of token:*PKCHAR pairs.  You can then have a separately updated registry of the tokens, which would have C, M, I, K,  present and which would be matched in your spec to the limitations.  As you introduce new tokens, you can update the registry.  If you go that path, I would suggest splitting class and channel so that they are independent tokens, but this is a stylistic suggestion only.
>
> My personal views only,
>
> regards,
>
> Ted Hardie
>
> On Mon, Sep 17, 2018 at 1:14 PM, Gaurav Jain <[e-mail]> wrote:
> Dear IETF URI review committee,
>
> Wi-Fi Alliance would like to submit a scheme registration request for your review and consideration. Attached with this email you should find the following:
> 1. Scheme registration request standalone document
> 2. Latest published Wi-Fi Alliance technical specification for Device Provisioning Protocol (https://www.wi-fi.org/file/device-provisioning-protocol-specification-v10) covering the details of the protocol, URI definition, syntax, URI format and associated information
>
> Since the registration request is for a ‘permanent’ registration, Wi-Fi Alliance has reviewed the requirements listed in section 3 (Requirements for Permanent Scheme Definitions) of RFC 7595 and have found that this registration request satisfies the listed requirements. As a next step, we request you to review this scheme registration request and provide your comments to us by Oct-15 2018.
>
> Thanks and regards,
>
> Gaurav Jain
> Senior Manager, Program Technology
> [phone] Office
> [phone] Mobile
>
> Wi-Fi Alliance
> www.wi-fi.org
> http://twitter.com/wifialliance
> www.facebook.com/wificertified
>
> Visit our blog, The Beacon, for Wi-Fi industry topics and trends
>
>
> _______________________________________________
> 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
>