Table of Contents

Project Candidates:


(1=high, 3=low)


FESL 2: Finish FESL 1. This involves building an API or DSL to allow simplified creation/management of policies. Try to get concensus on API vs DSL first.



Multi-Node Fedora High Availability. A node goes down, everything still works.



Multi-Node Fedora  Horizontal Scale-out. More objects/datastreams per node, and increased r/w throughput. For example, faster ingest.



Flexible Deployment. More than Tomcat and Jetty. For example, JBoss and Glassfish. OSGI would be helpful here. With multi-node, other questions arise (what kind of cloud, etc.)



Persistence/Storage Interoperability. Improved (higher level than "blob") storage abstraction. Content storage based on object attributes.



JCR/CMIS. Interop with other systems via standards.



Flexible/Extensible Event-Driven Workflow for Inside Services. Would facilitate asynch. storage among other things. Also experimenting with workflow on your site. E.g. add indexing, other hooks in your workflow without writing Java code. Could enable different out-of-box bundlings of Fedora.






Immutability in order to support higher concurrency. Needs more elaboration/thought.



Improving Disseminators. Cleanup of arcane XML. Make it easy for people. Separate REST API for disseminations.



Large Datastream Mediation. Why "E"/"R" datastreams are not enough: security. Backend retrieval could also just be improved. Fedora acting on behalf of user for access to backend storage.



OSGi. Increased developer productivity/happiness. Lowers the cost of experimentation. Could be done as part of another well-scoped activity (XML validation service/component, etc)



Web Admin (open standards-based, e.g. HTML5, but not SVG please). Value: maintainability. customizability. /* (tax-free, pick me!)



Refactor Fedora's configuration mechanisms to support easier deployment/installation



Rengineer installation procedures to provide "configure on first-run" style installation.



Refactor Fedora's tests and test machinery for efficiency and better coverage, as well as easier release



Engage Cargo in Fedora's build machinery to allow for better pre-release testing. See FCREPO-1061.



#trackbackRdf ($trackbackUtils.getContentIdentifier($page) $page.title $trackbackUtils.getPingUrl($page))
  • No labels