There has been a lot of talking about whether rel=”nofollow” is good or bad, useful or usless (check
Bitflux Blog :: Why nofollow is useless )
However I think that nofollow is quite useful for a different reason than preventing spam: preventing search engine results shown on “third party” websites mislead the search engines themselves.
Here is what I mean: Take for example g-metrics.com. Most of the pages of this site display results (actually the top-10 results) returned by google for specific queries. To do this, g-metrics.com makes use of Google Web APIs. So when google indexes one of these pages, it would increase the pagerank (or whatever ranking system a search engine is using) of a URL just because it had high enough pagerank to show up in the top 10 results…
However, using rel=”nofollow” can prevent this. (I have started using nofollow in g-metrics as soon as I read about it).
Similar cases could become a big problem for search engines, as similar APIs become more popular (Google, Technorati, del.icio.us, Amazon/Alexa, etc…) The way I see it, similar services may oblige or encourage developers to use nofollow when presenting “their” results.