Re: Voluntary (and non-) Standards (was: Support for Verifiable Claims)

> https://quod.lib.umich.edu/j/jep/3336451.0014.103?view=text;rgn=main#5.4

Thanks for the reference, I had forgotten about this document!  

> Distinguishing between the two approaches is important when talking about a new effort; if there isn't already wide adoption 
> and implementation of a proposal, it's necessary to do appropriate research and be realistic about the chances of success, 
> vs. the effort expended and opportunity cost. The reservations expressed about this proposal reflect that, I think.

I think that’s a very constructive proposal and maybe a way forward on the larger question of how to balance the views of users and implementers (or non-browser implementers) in W3C.  The Standardize Current Practice approach implied by the Rec Track Readiness criteria might be the smoothest and surest path to real standards that get implemented and used, but I see the point that sometimes it’s worth gambling that the Anticipatory Standards route could get there. The AB is discussing next week the Rec Track Readiness guidelines, Manu’s critique, and what we’ve learned from the VC discussions. I could imagine revisions that explicitly talk about when an anticipatory standards effort has a reasonable enough chance of success to be put on the Rec Track.  The number of members supporting it, saying they will participate, and promising to implement whatever comes out are at least part of the calculation.  Some other criteria, such as a thorough gap analysis of the existing standards landscape to justify the need for the anticipatory standard, and an explicit, plausible story for how the existence of WG could help bootstrap a viable ecosystem, seem reasonable to add.  I can also imagine that the Director would need to see concrete evidence that the ecosystem did indeed get booted up in order to approve advancing Anticipatory specs down the Rec track.


BTW, David, sorry for invoking the specter of the Spec That Must Not Be Named[1].  I certainly don’t want to start a substantive discussion of it, but if “those who do not learn from history are doomed to repeat it”, this is an episode I don’t wish to live through again.  The lesson I take away is to be extremely wary of standardizing something that doesn’t have buy-in from the people whose support is needed to make it a success.   

Apropos John Foliot’s point that You Know What is more successful than I think, it would be good for those proposing an Anticipatory Standard to define what “success” would look like up front so the team and AC can evaluate claims that it has succeeded. 

[1] Sorry for the oblique culture-specific reference to https://en.wikipedia.org/wiki/Lord_Voldemort, but I couldn’t help myself ☺

-----Original Message-----
From: "Nottingham, Mark" <mnotting@akamai.com>
Date: Thursday, December 8, 2016 at 4:45 PM
To: Adrian Hope-Bailie <adrian@hopebailie.com>, "Varn, Richard J" <rvarn@ets.org>, David Ezell <David_E3@verifone.com>, Nate Otto <nate@badgealliance.org>, "Stone, Matthew K" <matt.stone@pearson.com>, Manu Sporny <msporny@digitalbazaar.com>, Nathan George <nathan.george@evernym.com>, Kerri Lemoie <kerri@openworksgrp.com>, David Chadwick <d.w.chadwick@kent.ac.uk>, Jay Johnson <jay@qples.com>, Eric Korb <Eric.Korb@accreditrust.com>, Christopher Allen <ChristopherA@blockstream.com>, Phil Archer <phila@w3.org>, Gray Taylor <gtaylor@conexxus.org>, Bob Burke <bburke@kou.pn>, Linda Toth <ltoth@conexxus.org>, Drummond Reed <drummond@respectnetwork.com>
Cc: Tantek Çelik <tantek@cs.stanford.edu>, David Singer <singer@apple.com>, Chris Wilson <cwilso@google.com>, Michael Champion <Michael.Champion@microsoft.com>, "w3c-ac-forum@w3.org" <w3c-ac-forum@w3.org>, "public-webpayments-comments@w3.org" <public-webpayments-comments@w3.org>
Subject: Re: Voluntary (and non-) Standards (was: Support for Verifiable Claims)

    FWIW - 
    
    This experience isn't limited to the W3C. Carl Cargill calls it "Anticipatory Standardisation"; a good summary of the pitfalls is at:
      https://quod.lib.umich.edu/j/jep/3336451.0014.103?view=text;rgn=main#5.4

    ... and also his book, "Open Systems Standardisation."
    
    Personally, I refer to it as "hope-based standardisation." There is certainly a place for it, but making such an effort successful takes a considerable amount of very specialised work (or luck).
    
    The underlying question is whether you're attempting to use standardisation as a function of market consolidation (commoditisation), or market creation (effectively, R&D). 
    
    Distinguishing between the two approaches is important when talking about a new effort; if there isn't already wide adoption and implementation of a proposal, it's necessary to do appropriate research and be realistic about the chances of success, vs. the effort expended and opportunity cost. The reservations expressed about this proposal reflect that, I think.
    
    I've seen assertions go by in threads about this proposal along the lines of "that company has expressed interest in this, and they have fifty million users, therefore the potential market for this is incremented by fifty million." Or, more recently (to roughly paraphrase), "these states have expressed interest in digital licenses, and that problem can be solved by verifiable claims, therefore the potential market is larger."
    
    These things may be true, and it may even work out that what a future W3C WG specifies will actually be used in these ways, but it does NOT guarantee success. 
    
    This is one of the primary misperceptions that people have about standards work ("standardise it and they will come"), and is one of the reasons you're seeing such a strong reaction from the "old timers". If you want the discussion to move forward, I'd recommend that the advocates stop making assertions like this; it's a bit of a red flag.
    
    Hope this helps,
    
    
    
    
    > On 9 Dec. 2016, at 9:09 am, Michael Champion <michael.champion@microsoft.com> wrote:
    > 
    > > "long experience has shown in W3C that "if we standardize it, they will come as needed" is a generally false assertion."
    >  
    > Hmm, I was hoping someone would provide a couple of counter-examples.  The one usually cited is SVG, which gradually gained credibility and universal implementation long after the original Recommendation was published.   But that happened 15 years ago, and I can’t think of a more recent example.
    > John Foliot makes a good point about WCAG, but that is a *guideline* not a spec to be implemented.  And its success as a quasi-de jure standard comes after many years of relationship building, education, and successful use in the real world. 
    >  
    > One interesting data-based analysis is in https://blogs.windows.com/msedgedev/2016/02/03/2016-platform-priorities - “more than a third of web standards aren’t implemented by any of the most popular browsers.”   It would be interesting to research which supposed core web platform standards those are and how they got to be standardized without implementation support.  But this reality is what is driving the browser implementer community to insist on incubation before standardization.
    >  
    > More anecdotal evidence of notable failures of Rec-track work to be implemented, with painful consequences for W3C:
    >  
    > - XHTML2 is the poster child for an “aspirational” effort that had no support from implementers, and W3C’s persistence in trying to standardize an XML-based alternative to HTML drove HTML standards work out of W3C, to WHATWG.  W3C admitted error and tried to re-unite the communities a few years later, but the wound never healed and most technical work on HTML and DOM now happens in WHATWG (with W3C playing a useful role of publishing versions with broad and clear patent commitments).
    > https://en.wikipedia.org/wiki/WHATWG#History

    > https://www.wired.com/2009/07/the_w3c_buries_xhtml_2dot0_html_5_is_the_future_of_the_web/

    >  
    > - A more recent example is the HTML longdesc attribute, which became a W3C Recommendation over the objections of implementers, and even 2 years later AFAIK has little actual (correct) use on popular websites. For example, it’s not mentioned by caniuse.com, the de facto reference for which web “standards” actually work.  And it’s still not specified in the WHATWG version of HTML. Although it is not widely implemented or used, it still causes controversy and distraction.
    > http://caniuse.com/#search=longdesc

    > https://github.com/w3c/html/issues/507

    >  
    >  
    >  
    > From: Adrian Hope-Bailie <adrian@hopebailie.com>
    > Date: Thursday, December 8, 2016 at 1:18 PM
    > To: Tantek Çelik <tantek@cs.stanford.edu>
    > Cc: "Varn, Richard J" <rvarn@ets.org>, Gray Taylor <gtaylor@conexxus.org>, David Singer <singer@apple.com>, David Ezell <David_E3@verifone.com>, Michael Champion <Michael.Champion@microsoft.com>, Manu Sporny <msporny@digitalbazaar.com>, Nate Otto <nate@badgealliance.org>, "Stone, Matthew K" <matt.stone@pearson.com>, Chris Wilson <cwilso@google.com>, Mark Nottingham <mnotting@akamai.com>, "w3c-ac-forum@w3.org" <w3c-ac-forum@w3.org>, "public-webpayments-comments@w3.org" <public-webpayments-comments@w3.org>, Drummond Reed <drummond@respectnetwork.com>, Nathan George <nathan.george@evernym.com>, Kerri Lemoie <kerri@openworksgrp.com>, David Chadwick <d.w.chadwick@kent.ac.uk>, Eric Korb <Eric.Korb@accreditrust.com>, Christopher Allen <ChristopherA@blockstream.com>, Phil Archer <phila@w3.org>, Linda Toth <ltoth@conexxus.org>, Jay Johnson <jay@qples.com>, Bob Burke <bburke@kou.pn>
    > Subject: Re: Voluntary (and non-) Standards (was: Support for Verifiable Claims)
    > 
    >  
    > 
    > What I am still waiting for is a citation (or anything more than anecdotal evidence) for the following statement which seems to be the crux of all arguments I have heard against this work to date:
    > 
    > "long experience has shown in W3C that "if we standardize it, they will come as needed" is a generally false assertion."
    > 
    >  
    > 
    > On 8 December 2016 at 23:01, Tantek Çelik <tantek@cs.stanford.edu> wrote:
    > 
    >> On Thu, Dec 8, 2016 at 12:51 PM, Varn, Richard J <rvarn@ets.org> wrote:
    >> > I cannot say I much appreciate your discourteous tone.
    >> 
    >> Please don't misinterpret directness as discourteousness. I will
    >> interpret tone-policing of content criticisms as insecurity.
    >> 
    >> > Attached is the information on state of digital driver's licenses I got from the Internet in 3 minutes and 25 seconds.
    >> 
    >> Thanks I will take a look.
    >> 
    >> >  So there should be some greater use of citations by us
    >> 
    >> By everyone.
    >> 
    >> 
    >> > and some basic research by you
    >> 
    >> No.
    >> 
    >> This is part of the problem that Chris, Mike, David have pointed out.
    >> Any expectation from advocates that critics are supposed to do their
    >> own research is an unreasonable attitude of time-entitlement.
    >> 
    >> This is such a fundamentally flawed attitude that it further undercuts
    >> any faith in verifiable claims efforts.
    >> 
    >> Imagine if a recipient of a claim was told to "do some basic research"
    >> in order to verify it. It would be totally unacceptable as a protocol.
    >> 
    >> Tantek
    >> 
    >> 
    >> 
    >> > -----Original Message-----
    >> > From: Tantek Çelik [mailto:tantek@cs.stanford.edu]
    >> > Sent: Thursday, December 08, 2016 3:28 PM
    >> > To: Gray Taylor <gtaylor@conexxus.org>
    >> > Cc: singer@apple.com; David Ezell <David_E3@verifone.com>; Michael Champion <Michael.Champion@microsoft.com>; Manu Sporny <msporny@digitalbazaar.com>; Nate Otto <nate@badgealliance.org>; Stone, Matthew K <matt.stone@pearson.com>; Chris Wilson <cwilso@google.com>; Tantek Çelik <tantek@cs.stanford.edu>; Mark Nottingham <mnotting@akamai.com>; w3c-ac-forum@w3.org; public-webpayments-comments@w3.org; Varn, Richard J <rvarn@ets.org>; Drummond Reed <drummond@respectnetwork.com>; Nathan George <nathan.george@evernym.com>; Kerri Lemoie <kerri@openworksgrp.com>; David Chadwick <d.w.chadwick@kent.ac.uk>; Eric Korb <Eric.Korb@accreditrust.com>; Christopher Allen <ChristopherA@blockstream.com>; Phil Archer <phila@w3.org>; Linda Toth <ltoth@conexxus.org>; Jay Johnson <jay@qples.com>; Bob Burke <bburke@kou.pn>
    >> > Subject: Re: Voluntary (and non-) Standards (was: Support for Verifiable Claims)
    >> >
    >> > tl;dr: Who verifies the claims of the Verified Claims advocates?
    >> >
    >> > (motivation) If Verified Claims advocates can't be bothered to provide simple URL citations to verify their claims, why would anyone bother with anything more complex?
    >> >
    >> > (dogfooding) If you're not living breathing the behaviors you're advocating, why should anyone take advocations of (formalized versions
    >> > of) those behaviors seriously?
    >> >
    >> >
    >> > Longer:
    >> >
    >> >
    >> > Not picking on you in particular Gray, because this is an endemic problem that I have seen in pretty much all Verified Claims (CG/WG) discussions.
    >> >
    >> > Lots of claims made in the prose of such messages/emails, usually zero citations to verify those claims. Manu is the notable exception, he usually provides quite a few citations for his points in his emails.
    >> >
    >> > So just as an example:
    >> >
    >> >
    >> > On Wed, Dec 7, 2016 at 8:35 AM, Gray Taylor <gtaylor@conexxus.org> wrote:
    >> >> Interesting thread on legal standing.  Right now, 9 states are wrestling with putting verifiable drivers licenses on mobile devices (the paper artifacts we use today are eminently fraud prone - just ask any college student).
    >> >
    >> > Which 9 states? Citations to .gov sites that can be used to verify this "9 states" claim? Or a citation to a summary thereof itself with citations for the specific states?
    >> >
    >> >
    >> >>  In today's case, US State Department, DMV, Social Security Administration, County records, etc. all act as trusted service providers of the "paper and static ID" world; with great peril to the citizen as these artifacts can be stolen easily.  Their role won't change anytime soon.
    >> >
    >> > Presumably you're referring to passports, drivers licenses, social security cards, etc. and expecting (likely) that these examples are physically self-evident.
    >> >
    >> >
    >> >> Conexxus' feeling is that we don't proscribe legal purview of verifiable claims, but create an eco-system by which the "watchers" in today's existential data world can choose reliable new technologies to continue their mandated mission; and on a basis of NOT conveying unnecessary and static PII, which is the Achilles heel of our online existence.  So the intent is to provide control over our own identities as a first order.
    >> >
    >> > Could you provide a public Conexxus URL that describes this "eco-system" goal in more detail?
    >> >
    >> >
    >> >> If W3C creates a trusted environment framework, then the agencies will adopt them as a matter of public demand (IMHO this will be an escalating societal trend).
    >> >
    >> > This is a very shaky hypothesis, on multiple counts.
    >> >
    >> > First, agencies presumably adopt things without W3C involvement (e.g.
    >> > whatever they have adopted today).
    >> >
    >> > Second, what successful examples can you cite of W3C created standards involving trust (or anything else) that "agencies" subsequently adopted? Whether from public demand or other motivation. I have seen no evidence to support this "if ... then" hypothesis.
    >> >
    >> >
    >> >>  Each (global) jurisdiction will make its decision based on available technology and political aims v. the will of their people.
    >> >>
    >> >> Our retail industry does not want to know anything about you beyond "are you old enough to buy beer?" and can I capture the signature (read legal verification) of the TSP saying you are?  Certainly no business will stake their liquor license on a semi-trusted service provider, so the framework needs to authenticate the TSP as well.
    >> >
    >> > Presumably this is orthogonal or unrelated, as such businesses today seem to (anecdotally) only accept government issued IDs for "are you old enough". I would assume they will continue to do so, regardless of what tech happens to be in such IDs, and I'd doubt they'd accept non-govt issued IDs.
    >> >
    >> >
    >> >> So long opinion, short, if we build it, they will come as needed ...
    >> >
    >> > build yes, just standardize no. And this discussion is about creating a working group to create a standard.
    >> >
    >> > Specifically, long experience has shown in W3C that "if we standardize it, they will come as needed" is a generally false assertion.
    >> >
    >> > More TR RECs (https://www.w3.org/TR/) than not have failed to gain any serious broad traction (web browsers and servers implement a small subset of W3C RECs, not to mention IETF RFCs). The number of obsolete, abandoned, etc. W3C RECs and IETF RFCs greatly outnumbers those in modern use. I don't have exact numbers, merely from personal analysis.
    >> >
    >> >
    >> > <aside>
    >> >
    >> > The AB *is* working on a process for explicitly obsoleting abandoned RECs to start cleaning this up, in the hopes that eventually the RECs remaining are the ones that have actually be widely implemented, deployed, and are in use.
    >> >
    >> > We've started with a few examples to help us drive the necessary process changes:
    >> > * https://www.w3.org/wiki/AB/2016_Priorities#Specifications_to_obsolete

    >> >
    >> > </aside>
    >> >
    >> >
    >> >> who watches the watchers is the age-old question.
    >> >
    >> > who asks the claimers for citations for their claims?
    >> >
    >> > I'm going to keep asking for citations for claims until I see a cultural shift towards people who want Verified Claims as a technology providing URLs to substantiate their claims.
    >> >
    >> > I think everyone should adopt more of a [citation needed] practice, especially in this community.
    >> >
    >> > Tantek
    >> >
    >> >
    >> >> -----Original Message-----
    >> >> From: singer@apple.com [mailto:singer@apple.com]
    >> >> Sent: Tuesday, December 6, 2016 4:34 PM
    >> >> To: David Ezell <David_E3@VERIFONE.com>
    >> >> Cc: Michael Champion <Michael.Champion@microsoft.com>; Gray Taylor
    >> >> <gtaylor@conexxus.org>; Manu Sporny <msporny@digitalbazaar.com>; Nate
    >> >> Otto <nate@badgealliance.org>; Stone, Matthew K
    >> >> <matt.stone@pearson.com>; Chris Wilson <cwilso@google.com>; Tantek
    >> >> Çelik <tantek@cs.stanford.edu>; Mark Nottingham <mnotting@akamai.com>;
    >> >> w3c-ac-forum@w3.org; public-webpayments-comments@w3.org; Richard Varn
    >> >> <rvarn@ets.org>; Drummond Reed <drummond@respectnetwork.com>; Nathan
    >> >> George <nathan.george@evernym.com>; Kerri Lemoie
    >> >> <kerri@openworksgrp.com>; David Chadwick <d.w.chadwick@kent.ac.uk>;
    >> >> Eric Korb <Eric.Korb@accreditrust.com>; Christopher Allen
    >> >> <ChristopherA@blockstream.com>; Phil Archer <phila@w3.org>; Linda Toth
    >> >> <ltoth@conexxus.org>; Jay Johnson <jay@qples.com>; Bob Burke
    >> >> <bburke@kou.pn>
    >> >> Subject: Re: Voluntary (and non-) Standards (was: Support for
    >> >> Verifiable Claims)
    >> >>
    >> >>
    >> >>> On Dec 6, 2016, at 14:15 , David Ezell <David_E3@VERIFONE.com> wrote:
    >> >>>
    >> >>> To the first point, I’m not sure what you mean by non-voluntary standards organizations:  ... I’m not sure this non-voluntary distinction is worth fretting about.
    >> >>
    >> >> Some standards organizations (notably ITU) are the result of treaties, and some (including ITU) produce standards that can later have the force of law behind them.   “X’s sold or made available in country Y must comply with standard Z.”
    >> >>
    >> >> As you say, it’s not strongly relevant, except that in this field, some of the use cases for verifiable claims also intersect with legal requirements (e.g. being required to check the age of someone before selling them certain products). We easily back into the ‘quis custodiet custodies?’ problem if we’re not careful (who watches the watchers?) and wonder “who is recognized legally as being able to prove the age of a customer?”.
    >> >>
    >> >>
    >> >> David Singer
    >> >> Manager, Software Standards, Apple Inc.
    >> >>
    >> >
    >> > ________________________________
    >> >
    >> > This e-mail and any files transmitted with it may contain privileged or confidential information. It is solely for use by the individual for whom it is intended, even if addressed incorrectly. If you received this e-mail in error, please notify the sender; do not disclose, copy, distribute, or take any action in reliance on the contents of this information; and delete it from your system. Any other use of this e-mail is prohibited.
    >> >
    >> >
    >> > Thank you for your compliance.
    >> >
    >> > ________________________________
    >> 
    >  
    > 
    
    --
    Mark Nottingham    mnot@akamai.com    https://www.mnot.net/

Received on Friday, 9 December 2016 03:13:29 UTC