First up was John Law (I hope it was John Law - meandering shuttle bus from the airport meant I was a little late) talking about upcoming developments in Summon (and a couple of things I'd missed).
- Build your own Summon widget at http://jcu.summon.serialssolutions.com/widgets/ (although we've already got one in libguides).
- Discipline limits coming in the next release (limit to one or more areas of study - 'physics' for example
- Searching in ten languages now supported including CJK
- Tell Summon which database you prefer your results to come from
- Improvements to Ebook Access
- Releasing new linkers (openurl2publisher algorithms) dynamically - ie as soon as developed/fixed rather than as batch monthly updates
- improved non-roman char handling
- Bigger suite of stats - including a referrer report (Finally we can see how many full text hits come where (Google Scholar for example) and a 'item with no full text' YAY - SFX in 2007 had these.
- A new iteration of the One-Click helper screen
There was a little discussion about Summon ranking mechanism where John provided some insights in how complex an algorithm you had to have when you ranking 80 different formats with metadata ranging from the bare bones of a citation up to the actual full content of an ebook. It made me think about how some people who scratch their heads at the order of a Summon results list and say they wish it wise like Google. In Google you are looking for one thing, and if it appears in the first five hits you're happy, you don't scroll down to record 15 and wonder why it wasn't record number 2 - but you do in Summon because your needs are different to a Google search, you don't have a simple single information need, you have a complex expanding requirement that you heuristically refine as you as you go.
Then someone jabbed me and said it was time for a break.