[apps-discuss] Usability of draft-nottingham-safe-hint (was: Call For Adoption: draft-nottingham-safe-hint)

Thinking about it further, this raises a great point: As currently
defined, "Prefer: safe" is actually quite useless because the
specification does not provide any basis upon which to determine
whether content is "objectionable" or not. The challenge is that it
attempts to address the problem in the wrong way.

Consider a school. It rolls out a bunch of tablets with the "Prefer:
safe" boolean option set. Two different students go to two different
websites, each of which say they support "Prefer: safe" but have two
entirely different interpretations of what the currently undefined
concept of "objectionable content" means. The result is that the
school ends up having to apply it's own filtering anyway, despite the
presence of the "Prefer: safe" option, making the use of "Prefer:
safe" pointless and meaningless beyond making someone feel good.

The problem lies in the fact that "safe" is not bound to any clear,
testable condition. For instance, we could say that a server honoring
the "Prefer: safe" preference SHOULD NOT return any images, videos or
textual content depicting sexually explicit or sexually related
subjects. That would be a testable behavior that could be expected to
be remain consistent across implementations. Without such a clear
definition of what "safe" means, the spec, as currently written, is
largely meaningless.

- James

---------- Forwarded message ----------
From: Karl Dubost <karl@la-grange.net>
Date: Wed, Jul 23, 2014 at 6:19 PM
Subject: Re: [apps-discuss] Call For Adoption: draft-nottingham-safe-hint
To: "Murray S. Kucherawy" <superuser@gmail.com>
Cc: IETF Apps Discuss <apps-discuss@ietf.org>

Le 22 juil. 2014 à 01:28, Murray S. Kucherawy <superuser@gmail.com> a écrit :
> This note begins a Call For Adoption for the above document to be adopted as an APPSAWG working group item.  The Call ends on August 8, 2014.

potential for ratholes discussions.
a "safe" preference is strongly attached to a cultural context.
Cultural contexts are related to groups and often physical access to
places with a notion of sealed environments.

A magazine stand will put some publications into a dedicated
environment (hidden, special room, etc) according to the law of the
country (or the province). The mechanism of enforcement is often done
through the barrier of a human person controlling the access. Same
thing for alcohol, tobacco, amusement park (height), etc.

Putting a label on the browser side is indeed a way to advertise that
you do not want to receive some type of content, but the issue is that
"safe" has no meaning with regards to the Web because the cultural
concepts associated with the word "safe" is highly dependent on the
community and geographical location of the person using the browser.

This document is likely to lead to long discussions, will not solve
anything in a satisfying way, will be abused in all the ways you can
imagine. It will be even a potential source of additional tracking for
ads businesses along "This person is probably religious or a child,
let send targeted ads."

--
Karl Dubost 🐄
http://www.la-grange.net/karl/

_______________________________________________
apps-discuss mailing list
apps-discuss@ietf.org
https://www.ietf.org/mailman/listinfo/apps-discuss