The library community has deployed several niche protocols. Most were developed before the web, and perhaps need to be updated to be more 'webby' if they are actually to be more widely used. Certainly, in their current state they are unlikely to be used outside of the library community whatever about within.
Two protocols which are more 'webby' are the Open Archives Initiative Protocol for Metadata Harvesting and SRW/U (Search and Retrieve on the Web and Search and Retrieve over a URL). The latter recasts the functionality of Z39.50 in the idiom of web services; the former is a relatively new initiative which emerged within a web idiom from the start.
Rob Sanderson, from Liverpool University, and my colleagues Jeff Young and Ralph Levan, have written an article about these two protocols and how they can work together.
These two protocols look set to form a strong, interoperable backbone for the distribution and remote discovery of data for many years to come, so it is worth looking at how they can be used together advantageously. While some combinations are immediately obvious, there are other architectures that have not been discussed and are worthy of notice. It is also interesting to look, briefly, at how the two protocols handle the same fundamental requirements. [SRW/U with OAI: Expected and Unexpected Synergies]Incidentally, one of the applications they describe is the OCLC Research Publications Repository which has been built from several tools stitched together with these light-weight protocols.