In case you're wondering, I finally managed to clear my Bloglines backlog, including the published literature, D-Lib and Ariadne. I found two very different, but complementary views on how service standards and standard interfaces can enable an enhanced scholarly workflow or other advanced combinations of services.
In Serving Services in Web 2.0 (Ariadne issue 47, April 2006, ISSN 1361-3200), Theo van Veen of the Koninklijke Bibliotheek in the Netherlands explores and explains some fundamental concepts of Service-Oriented Architecture and standard service interfaces.
In this article I discuss the ingredients that enable users to benefit from a Service-Oriented Architecture (SOA) by combining services according to their preferences. ... This concept is an extrapolation of the use of OpenURL and goes beyond linking to an appropriate copy. Publishing and formalising these service descriptions lowers the barrier for users wishing to build their own knowledge base, makes it fun to integrate services and will contribute to the standardisation of existing non-standard services.
In An Interoperable Fabric for Scholarly Value Chains (D-Lib, October 2006, Volume 12 Number 10, ISSN 1082-9873), Herbert Van de Sompel, Carl Lagoze et al explore how you can build services using an interoperable network of digital object repositories
This article describes an interoperability fabric among a wide variety of heterogeneous repositories holding managed collections of scholarly digital objects. These digital objects are considered units of scholarly communication, and scholarly communication is seen as a global, cross-repository workflow. The proposed interoperability fabric includes a shared data model to represent digital objects, a common format to serialize those objects into network-transportable surrogates, three core repository interfaces that support surrogates (obtain, harvest, put) and some shared infrastructure. This article also describes an experiment implementing an overlay journal in which this interoperability fabric was tested across four different repository architectures (aDORe, arXiv, DSpace, Fedora).