[microformats-discuss] hreview - aggregation

[Ryan King]

On Aug 27, 2005, at 4:25 PM, Stephen Downes wrote:

> Hiya,
>
> I have been reviewing the hreview specification http:// 
> microformats.org/wiki/hreview which on the whole looks pretty solid.

Great.

> I am wondering whether people have considered the expression of  
> aggregated reviews and/or ratings.

I don't think that's been discussed yet. Honestly, just getting  
adoption of hreview'ed individual reviews is enough for now (IMHO).

> In such a case, the ratings values themselves would be collected  
> and averaged, presented as a single value.

Granted, people will always want to present summaries of individual  
reviews. However, I'm not sure that its important that this data be  
machine consumable. At least, I wouldn't put it as a high priority.  
So long as the individual reviews are available for consumption, the  
summary data is redundant. A UA could just do the aggregation  
themselves.

Having said that, this issue will probably come up again and we'll  
probably need to address it sometime in the future. So perhaps you  
could add these ideas to http://microformats.org/wiki/hreview- 
brainstorming so that they don't get lost.

thanks,
ryan

> This, from what I can see thus far, could be accomplished with the  
> existing spec, with two changes:
>   - 'rating' would allow real numbers (eg., 3.5) and not only integers
>   - an additional element, 'number' or 'n' would indicate the  
> number of ratings collected to produce the average
>
> An aggregation would then (optionally) list individual ratings,  
> along with the identity of the evaluator.
>
> Since I *will* be doing this in the field of learning resources, it  
> would be helpful to have a canonical vocabulary in which to express  
> it.