13
edits
Changes
no edit summary
The open source (AGPL3) [http://archivematica.org Archivematica] digital preservation system uses a micro-services architecture to integrate a suite of Linux utilities into workflow pipelines. It is designed as a backend tool for archivists and librarians managing digital collections and digital preservation responsibilities. We use Google Gearman for job scheduling and load balancing as well as Django (python) for a web-based administration interface that monitors and controls the processing of files in the pipelines. The system creates standards-compliant (e.g. METS, PREMIS, Bagit) archival packages as well as a registry interface to monitor format policies. This system is designed to provide the technical component for ISO 14721 (OAIS) and ISO 16363 (TRAC) compliant Trusthworthy Digital Repositories. The recent 0.8 release is the last alpha. Over winter 2012 we are continuing with scalability testing and tuning, adding ElasticSearch indexing, SWORD deposit support, interfaces for Dspace, ContentDM, XTF; all for inclusion in the 0.9-beta release sometime in Spring 2012. The presentation will give a quick demo of Archivematica's features as well as discuss technical architecture, APIs, development roadmap, user base, community building, project management, etc.
== Virtual Integrated Search - on-the-fly merging of relevancy ranked searches ==
* Mads Villadsen, The State and University Library Denmark, mv@statsbiblioteket.dk
What do you do when you have an integrated search system and the users want data at the article level? What we did was to try and get the data from the publishers - and when that failed we went with Summon for the article data while keeping our bibliographic records (and more) in our own system.
So how’s that working out for us?
We didn’t want to give up on our overall goal of having a single unified result set which meant we had to do something out of the ordinary.
We struck a deal with Serials Solutions that allowed us to apply our technical know-how and sprinkle fairy dust on our queries thereby achieving a proper relevancy ranked merging of results from our own index with the results from Summon. We gave a lightning talk about some of these ideas at last year's code4lib.
We have been running this "Virtual Integrated Search" in production since August and the end users haven't come at us with their pitch forks yet so we assume they are still able to find what they are looking for.
Just to be sure we will be performing a usability test in November 2011 that will hopefully guide our future development.
I will cover what goes into making fairy dust ("how it works", "what doesn't work") as well as some of the results from the usability test ("does it actually work?").
http://www.statsbiblioteket.dk/search/
[[Category: Code4Lib2012]]