A distributed search portal of common sources of ISBN numbers, with permanent caching of results. To provide a free software, open-source, interface for ISBN retrieval using HTML, SQL or XML to be independent of any toolkits or software.
There are significant changes throughout the entire package for v0.3!
The distributed framework is now operational - each site has full access to all ISBN records retrieved. This is the single largest development in the project and enables the formation of a complete index of all ISBN records - available to everyone and anyone, without restriction or limitations.
There is comprehensive documentation now included in the package and on the sourceforge project pages.
Each isbnsearch site is configured as either a portal or indexserver. Portals are linked to indexservers in a distributed framework that keeps track of the most common ranges of ISBN data in each database. Indexservers collate the range information from many portals to build a predictive model of which ISBN records can be retrieved from which portals.
Records that are not found locally on any one database are retrieved by using the model to refer any missing query to the portal most likely to have already stored the data. The portal will output an empty file if the record is still pending and the search will continue in the background using an automated Z39.50 cron task. Users can re-try any pending query or view the status of pending records at any time.
With enough sites, the intention is to cover all 10 digit ISBN records before 2007 when the model will be updated to start storing the new 13 digit format.
So the target now is to implement as many servers as possible - if you have 250Mb or more of webspace available, please download the package and join the mailing list!
SourceForge project:
http://sourceforge.net/projects/isbnsearch/
0.3-1 release:
http://sourceforge.net/project/showfiles.php?group_id=90779&release_id=193155