Hi Gannon, [Sorry for the delay, your mail accidentally skipped my inbox!]
> My question can be rephrased thus: Does the theoretical size of the target > audience for a distributed affordance matter ? The audience size doesn't matter, as each user has personal preferences. The idea is that the client has an extra piece of intelligence that adds the actions the user cares about. For instance, if you prefer to see movies in a certain theatre and always buy books through Barnes & Noble, then you will "bookmark" those actions. When your client (browser) then arrives on a page with a movie or book, it will automatically insert links to buy tickets for that theatre or order through Barnes & Noble, instead of suggesting Netflix, Amazon or iTunes. As the processing happens on the client side, we can scale without impacting the infrastructure. > It seems to me that "distributed" entails that one size fits does not fit all > clients, that constrains the Open World … Exactly! But this is why the client performs the adaptation: it's personalized and scalable in the number of clients. Is this the answer you're looking for? Additionally, the diagram in http://ruben.verborgh.org/publications/verborgh_wsrest_2013/ might clarify things. Best, Ruben
