Re: Social use case ideas

Hi Uldis, (and Jodi, as I see she started making changes on the wiki about this new "social" cluster)


Quick feedback on selected points...

  
> 1) Social annotation
>
>[...]
>
> 2) Social recommendations
>
> Recommend books based on social data (annotations, usage data)
>    - use user activity (how many people annotated a book, looked at a
> book, read it, ...) for determining its relevance to the user and in
> recommending books to look at.



I find 1) and 2) very interesting: in fact is a continuation of the other, isn't it?


> How does Google Books rank book search results?


I found this: http://www.theatlantic.com/technology/archive/2010/11/inside-the-google-books-algorithm/65422/
It's not really clear how precisely they do it, and where they find the relevant usage data (except for "web search frequency"; do they fetch data from services like Worldcat? Amazon?). But it points at the direction you hinted at yourself regarding using user activity.


> 3) Location/ownership info
>
> An open publication catalog + information about where these
> publications can be found.
>
> To add a social aspect, let users of this service indicate what
> publications they have and if / on what conditions they would lend
> these books. As a result participants would not only be able to find
> books in nearby collections (presumably covered by the Use Case Find
> materials in the closest physical collection) but could also enable
> peer-to-peer book-swapping (though would require a critical mass of
> users from any given location before it becomes useful for people from
> that location).


I'm not sure I understand this use case: who are the "users"/"participants" who hold publications and could swap them? End users or libraries themselves?


> Related: Open Library (linked data via the Use Case Open Library Data)
> could provide information about publications.
>
> 4) Add social features to ILSs
>
> Simple and not-so-interesting social features are buttons for sharing
> an item on Facebook, Twitter, etc. Not that interesting to us as they
> are already present on many sites (and probably on some ILSs) and
> there is probably little use for linked data there.


If the idea here is to share book data a la Facebook OpenGraph API, I think this is very linked data-compatible :-)
In fact I would see such scenario as a better motivation for putting the "SEO" case [1] in your social cluster [2]. For now I can't easily see the reason for putting it there. There's not much "social" stuff in that SEO case!

Cheers,

Antoine

[1] http://www.w3.org/2005/Incubator/lld/wiki/Use_Case_SEO
[2]http://www.w3.org/2005/Incubator/lld/wiki/UseCases#Cluster:_Social_Uses

Received on Wednesday, 5 January 2011 16:10:50 UTC