I am testing how it is possible to aggregate statistics on the same publication via its multiple IDs. For instance, this paper of mine has DOI, mendeleyID and a URL: http://total-impact.org/collection/LTtNmS
When I retrieve stats with TotalImpact I see that e.g., the number of CiteULike bookmarks actually refers to the same paper, no matter if I use DOI or mendeleyID, while for the number of readers on Mendeley, it is different. Also, later topsy should tell that for the URL there is 1 tweet, and I suppose for DOI and mendeley ID it'll still show 0 tweets.
My question is, did you have any plans of supporting multiple IDs and what you thought about this?
My impression was that if I know that my pub has multiple IDs, I'd like to provide all of them (and group?) and then get an aggregate stats. For the aggregate I'd need to know if I should sum up metric values for various IDs or do not sum up (does it depend on metric, or, rather, the way it is obtained via some API?) and show aggregate stats.
Do you think it could be useful for TI API to make it possible to pass multiple IDs and get overall stats? Is there such functionality? Can we (I and Eugene) implement it or it is better to keep such things inside PeerEvaluation?