Well, just one example is searching twitter. If you want to search for
tweets that mention a particular url, you need all of the short urls
used for that destination url. Obviously there are more url shorteners
than just
bit.ly but
bit.ly is the most used. Without having a way of
getting all user hashes for a global hash, you would need to index
every tweet with a
bit.ly link, store it in your own database and run
it through the
bit.ly /expand api method so that you can have a
references of all
bit.ly links used for a single destination url.
Doing something like that is a serious undertaking and extremely
inefficient, whereas being able to request all of the user hashes for
a global hash of a destination url *when it's needed* saves everyone a
lot of work and a lot of api calls.
There are certainly other examples where an api method like this is
useful, but they all pretty much boil down to the same thing: being
able to aggregate data yourself.
Thanks for listening and I hope we can see something like this in the
api soon!