Re: web architecture and safe content

On Sun, Oct 11, 2009 at 11:58 AM, Karl Dubost <karl@la-grange.net> wrote:
>
> Le 11 oct. 2009 à 08:56, Mukul Gandhi a écrit :
>>
>> I feel, a preferences like, "SafeSearch Filtering" must be part of W3C
>> web architecture (or perhaps some kind of search engine guidelines).
>
> What is "Safe" is very dependent on one's culture.
> Example: a singer nipple on stage made a big mess in USA when it would not
> have raised any eyebrows in some european countries.
> Many examples can be chosen for language level, political ideas, etc.
>
> It is then almost impossible to label photos in a binary safe or not safe
> way. There are many types of safe or non safe which depends on the
> individual and one's community. Maybe a better option would be to develop
> community trust mechanisms. One's would rely on his/her community notion of
> safe URIs (system of shared bookmarking etc.) That would make an interesting
> semantic Web project.
>
> POWDER is an attempt at that, labeling web pages.
> see http://www.w3.org/TR/2009/NOTE-powder-primer-20090901/

Yup. POWDER adds to RDF's expressivity one small but critical thing
that was missing in the older PICS standard; a way to say that a
description applies to anything whose URI matches some regular
expression.

A pre-POWDER format, RDF-WCL is available at
http://www.playboy.com/labels.rdf and gives the general idea, in this
case a label for Web content on playboy.com summarised as claiming
"Exposed breasts; Bare buttocks; Erotica; No violence; Mild
expletives; No potentially harmful activities; No user-generated
content; This material appears in an artistic context;" ...

Youou might dispute that last point, ... but to me that *is* the
point. Different parties can make broadly interoperable claims using
the same system, even to the extent of agreeing or disagreeing on the
rating vocabulary.

W3C has an old statement of principles document on just this,
http://www.w3.org/TR/NOTE-PICS-Statement and which discusses many of
the points bound to re-emerge when the use of RDF or POWDER is
considered for such applications.

cheers,

Dan

Received on Sunday, 11 October 2009 11:03:04 UTC