Wednesday 6 August 2008

Sprint 3 - Washup and planning

A few months ago we started to experiment with scrum, sprints and backlogs in an attempt to formalise our agile development processes. I thought it might be fun to try and keep the blog updated every couple of weeks with what's moving in the k-int world.. so here goes...

Hana's done some great work on the zero functionality release for the JISC Transcoder Project and it's basically finished. I'm tasked with uploading this to the Amazon elastic cloud this week whilst hana pushes on with having a go at SCORM 2004 to IMS CP 1.1 trascoding. Hana has also spent a couple of days fixing a drag and drop bug in the vocabulary editor application and getting reload installed for some testing. We're on the lookout for an SCORM 2004 content to test with (Especially older BBC JAM content).


Liams been doing loads of work refactoring the import and export modules in Vocabulary bank to make sure we can ingest and export both ZThes and SKOS and crosswalk between the two and in preparation for creating some additional export formats for specialist educational / eLearning applications. Liam's also managed to create an installer for the new release of OpenCRM, our open source third sector call management / CRM application to go off to Sheffield Advice Link. We've been told that the latest released vocabulary editor can import the full 7923 terms from the IPSV Integrated Public Sector Vocabulary in 17 seconds, which isn't bad going, considering the amount of cross referencing and revision management thats going on under the hood.



I've been on holiday ;) but whilst I've been away I've made huge steps forward on a stand-alone OAI/ORE server that we can use to replace all the custom OAI servers we have dotted around the place. With the new server you just point the app at a JDBC data source and it does introspection to discover the DB schema, detect primary keys and timestamps, a few clicks later and you have an OAI/ORE feed for your database. The same infrastructure can be used in online or batch mode (Batch mode detecting changed records with checksums rather than timestamps, useful for normalised schemas where changes might not update datestamps). Initially this is for an update to the Peoples network discover application, but I've high hopes it will be useful in a wide range of cultural and governmental settings. This should give us an instant linked-data capability for projects like peoples network and other related metadata aggregator projects we're involved with at the moment.

I've also (To my eternal shame in the office) started to do some real work on the JZKit documentation and the maven auto generated site, to make it easier for users of the toolkit to get to grips and diagnose problems. There seem to be a huge number of Z3950 projects in the offing at the moment, for a protocol thats supposedly legacy. I think thats great!

Rob's made huge steps forward with the assisted tagging tool which is being used (in anger by many users) to create descriptive metatdata for literally thousands of learning resources. The new tagging tool seamlessly integrates with the bank vocabulary service so we can update the vocabularies used without updating the tool, and the new assisted tagging feature makes it easy for users to quickly tag resources without needed an in-depth knowledge of curriculum and administrative structures.

Knowledge Integration Ltd