RefDB has changed a lot since the last mention of it here. From the latest release news on fm:

"The PHP Web interface supports live links for keywords, authors, and periodicals which are displayed as "tag clouds". Automatic format detection from local files is done, and has a type-sensitive form for editing data. RefDB implements all SRU operations (explain, searchRetrieve, scan) with MODS output, and conforms to CQL Level 2. Namespaced XML output allows processing of schema-based TEI P5 and DocBook V.5.0 documents. Raw bibliographies were added. Searching for and styling of less-common fields was improved."

The source, perl client, and sru support code are attached.

Crosswalking is a concise book for library programmers who want to learn to use MARC4J to process bibliographic data. MARC4J is an open source software library for working with MARC, MARCXML and related bibliographic standards in Java. The library is designed to bridge the gap between MARC and XML.

It is divided into the following chapters:

Chapter 1, Reading Data

Chapter 1 provides a short introduction about MARC formats and then focuses on reading MARC and MARCXML data. This chapter also explains how to create and update records and it demonstrates how to pre-process the input to convert MODS to MARC.

Chapter 2, Writing Data

Chapter 2 concentrates on the details of writing MARC and MARCXML data and how to post-process the output to convert MARC to MODS.

Chapter 3, MARC4J and JAXP

Chapter 3 explores integration with some important Java XML API's including JAXP, SAX and DOM. It demonstrates how to write the result to a DOM document, how to format XML output using a dedicated XML serializer, how to build pipelines using XSLT and how to use the SAX interface as an alternative to XSLT.

Chapter 4, Indexing with Lucene

Chapter 4 concentrates on indexing and searching MARC data with Apache Lucene using the MARC4J Lucene API.

Chapter 5, Putting It All Together

Chapter 5 focuses on building an SRU Search/Retrieve Web application using the various MARC4J interfaces and classes to process MARC data and using Lucene for indexing and searching.

Appendix A, MARC4J API Summary

Appendix A provides a summary of the core MARC4J interfaces and classes.

Appendix B, Command-line Reference

Appendix B documents the command-line programs included in the MARC4J API.

This book provides useful information for both developers learning about MARC4J for the first time and developers returning for reference and more advanced material. The chapters provide many reusable examples, while appendixes provide a reference to the API and the command-line utilities.

Crosswalking is published through lulu.com.

Visit lulu.com for more information.

refbase-0.9.0 adds support for unAPI, COinS, and a SRU/W web-service:

refbase is a web-based bibliographic manager for scientific literature, references and citations. A new release of the refbase package is available from SourceForge.

This release offers major function enhancements and bugfixes. Batch import from various bibliographic formats (including BibTeX, Endnote, RIS, ISI and MODS XML) is now supported, as is import from a PubMed ID. An OpenDocument Spreadsheet for use with OpenOffice.org can be exported and formatted citation lists can be generated as HTML, RTF, PDF, or LaTeX. A SRU/W service as well as support for unAPI, OpenURL and COinS metadata have been added. These allow the data to be used by the next generation of bibliographic clients. A new command line client is also included.

See Feature Highlights for an overview of the main features of this package.

Please see the README file for an overview of refbase, what has changed in this release, and how to install or update refbase. The README is available in the package and online.

For more information about the refbase project and pointers to working examples of refbase (including freely accessible demo databases) please visit the refbase home page.

From fm: "A bug that caused too many temporary image files to be written was fixed. Searching for multiple ISBN values using z39.50 was fixed. A layout bug for paragraphs in the Fancy template was fixed. IBS searches now reverse the author name. "Xbox 360" was added to the Amazon.com game searches. Deleting and re-adding a field by the same name no longer crashes. Default fields are no longer recreated when adding search results. The Spanish translation wasu pdated."

Visit the Tellico site for more info.

From freshmeat: "Major bugfixes: A potential data loss bug that occurred when changing the config option for image location and selecting File->Save As was fixed."

Anyone know if this can be built on OSX?

anon writes: "A new version of refbase, a web-based bibliographic database, has been released. The new version adds export to MODS XML/bibtex/endnote, Endnote import, RSS, and more."