how to do aggregate reviews - Re: [uf-discuss] hReview feedback

[Tantek Ç elik]

On 1/20/06 2:21 PM, "Benjamin Carlyle" <benjamincarlyle at optusnet.com.au>
wrote:

> On Wed, 2006-01-18 at 12:13 -0600, Paul Bryson wrote:
>> "Ryan King" wrote...
>>> On Jan 17, 2006, at 10:42 PM, Paul Bryson wrote:
>>>> A good portion of the numerical ratings on the internet are  aggregates
>>>> of
>>>> many people voting, usually without their own detailed reviews, so
>>>> getting a
>>>> float value for the rating would be pretty likely.
>>> I think this missed the point. Its those individual votes which are
>>> candidates for hreview, not the aggregate data.
>> I would have thought that they would both be canidates.  One is as common as
>> the other, and there is so much overlap that it seems wasteful to me to
>> ignore aggregate data.
> 
> I think it may be especially important to be able to mark up aggregate
> ratings as well as individual ratings, given that the aggregate may be
> of data not available to the web as individual ratings (think Amazon).
> It does seem likely, however, that the hReview format as written is not
> a good match to aggregates. It is designed to include text descriptions
> as well as a simle rating.

I am tending to lean towards this point of view as well.

Perhaps the right thing to do, for the folks interested in representing
aggregate reviews, would be to start documenting some examples, and then we
could discern how close to "normal" reviews the aggregates were, and if they
commonly had fewer/more fields (e.g. number of reviews that went into the
aggregate etc.)

I think John Panzer started this particular thread.

John, consider starting this page and documenting current examples of
aggregate reviews on the web so we can have some facts to reason from:

 http://microformats.org/wiki/aggregate-review-examples

Thanks,

Tantek