Video from a conference presentation on Library APIs
Richard Wallis presented at the code4lib conference in Athens, Georgia, last week, talking to the title "Library APIs abound". Read on for a link to his slides... and a video of the whole presentation.
Richard's presentation to the 2007 code4lib conference explored;
"From Z39.50 to xISBN, they share the limitation of providing a single stream of data from a single source.
How to add value to data from one source with relevant data from another, and how do you orchestrate that interaction in a scalable way?
A review and practical demonstration of augmentation APIs and their orchestration in a way that would make those used to Unix Pipes principles, feel at home."
Richard's Powerpoint presentation is available here, and a video captured by the conference organisers is also available for viewing here (135Mb, 20 minute .wmv file).
Z3950 Single Stream from a single source?
Mr Wallis, I'm surprised that you would suggest such a thing? Z9.50 says nothing about it's source, and it's completely agnostic about it's payload. Indeed, we've found that actually, Z39.50 is far superior to SRW for meta-search because the database abstraction allows source selection beyond the scope of what SRW will allow (Although we did argue long and hard in the definition of SRW about the presence of a list if databases to search). The bigger problem is with the lack of static identifiers coming back from all search services. Once known items have stable identifiers, augmentation becomes possible on many different levels, and not just within one host container. Indeed, orchestration is the key issue, the question then becomes, should the mechanism of that orchestration belong to the community or a single vendor?