Re: trust models and AWWW (was Re: FYI, tag election links)

On 18 December 2012 16:56, John Kemp <john@jkemp.net> wrote:

> On Dec 16, 2012, at 7:59 PM, Larry Masinter wrote:
>
> > Melvin,
> >
> > While the TAG might resist picking winners and losers in an unclear
> technology battle as a short-term strategy, I think AWWW suffers because it
> assumes trust too many places.
> >
> > For updating the AWWW we have, I’ve been thinking somewhat the opposite:
> that we need at least a model and a framework for talking about trust, and
> to encompass the different ways in which trust affects architecture.
>
> What is wrong with the following as a basis for talking about trust?:
>
> "The World Wide Web (WWW, or simply Web) is an information space in which
> the items of interest, referred to as resources, are identified by global
> identifiers called Uniform Resource Identifiers (URI).
> Examples such as the following travel scenario are used throughout this
> document to illustrate typical behavior of Web agents—people or software
> acting on this information space. A user agent acts on behalf of a user.
> Software agents include servers, proxies, spiders, browsers, and multimedia
> players."
>
> I think that you do need to describe the agents at play on the Web. How
> have the agents described in AWWW changed in the "web of apps"? Or have
> they changed at all? What would better-describe their interactions these
> days than the interaction section of AWWW?
>
> I did try this a few years ago myself, by the way:
> http://www.w3.org/2001/tag/2010/03/web-apps-taxonomy/web-apps-taxonomy.htmleven mentions "trust" explicitly.
>
> I do think that any new edition of AWWW would have to address whether the
> agents and the interactions between them are fundamentally changed, or
> whether the "sea of Javascript" is just a layer on the bedrock of the same
> set of agents interacting roughly in ways already documented by AWWW volume
> 1.
>
> Once you understood what entities are at work on the Web and how they
> might reasonably interact, you may be able to understand what properties
> are needed for each of them to be able to decide to trust each other or not.
>

Thanks for sharing the doc.  This makes sense.  Agents on the web are
exposed to claims either in natural language, or machine readable form.  In
the most cases, those claims can be taken at face value.  In certain
situations, it may be beneficial to verify some claims to gain additional
confidence (even though there is an overhead associated with the
verification process).  But also bear in mind that trust is subjective, and
can never be perfect.  A neutral framework that could take this into
account, and also introduce scalable Web of Trust, could be indeed
beneficial.


>
> Regards,
>
> JohnK
>
>