[uf-discuss] Exploratory discussion: content rating

[Dave Cardwell]

Andreas Haugstrup Pedersen wrote:
> On Thu, 27 Jul 2006 14:23:05 +0100, Drew McLellan wrote
>> From a discussion in #microformats trying to brainstorm around a  
>> gravatar.com-like system based on distributed hCards, we stumbled  
>> across perhaps a simpler problem to solve.
>>
>> All content published on the web is unrated (in terms of nudity,  
>> violence, profanity etc) by default. There is currently no *simple*  
>> method for authors to express a rating for their content so that  
>> visitors may choose to filter on their level of comfort.
> 
> The problem with ratings is not that they're hard to write (or rather, that 
> problem is easily solved). The problem is determining what ratings mean. 
> NSFW in one location can be very different than NSFW in the building next 
> door. An R rating may imply that anyone under 17 shouldn't be watching this 
> (according to the MPAA), but it can also mean that anyone over 13 should be 
> fine. And you have variations based on the actual content matter. MPAA is 
> hard on sex, but soft on violence, European rating systems may be different.
> 
> The cultural issues involved in creating a ratings system that actually 
> *work* in practice, across borders is really really hard. You need to solve 
> that problem before a microformat for ratings makes any sense.
> 
> IMO, of course.
> 
> --
> <http://www.solitude.dk>


The differing personal and cultural views of "morality" were discussed 
on IRC.  It was suggested that a microformat could instead indicate what 
sort of content was represented in the element.

For example, class="content-nudity content-profanity".  I don't think it 
is unreasonable to collect a number of such "moral" (for want of a 
better word) red flags (eg. violence, sex, and so on) that are fairly 
universal.  Obviously things like "content-tiananmen-square" are 
entering the realm of metadata which I believe is outside of the scope 
of what this microformat should be.

If we look at an existing example of content filtering implementation 
such as ICRA (http://www.icra.org) this is obviously something that is 
desirable to content generators (ICRA's corporate members include AOL, 
Microsoft...), content carries (...T-Online, British Telecom), and 
every-day content consumers wishing to protect themselves/their 
children/employees/etc from unsavoury content.  I believe this is beyond 
the scope of existing microformats like rel-tag and hreview.

Where I believe ICRA falls down is the complexity of its implementation. 
  If a uF for content "rating" is desirable, and can designed with 
simplicity in mind, I think it's uses are varied and extensive.  Of 
course, further exploratory discussion is required before we can be sure.


Best wishes,
Dave Cardwell ~ http://davecardwell.co.uk/