First of all, thanks to everyone who came to the VuFind Summit this year. I appreciate the investment of time and energy into the project, and it was good to talk to so many people
in person. It would have been nice to have a little more time to make sure I had one-on-one time with everybody, but I think I did reasonably well given the schedule -- apologies to anyone who felt neglected! I also get the impression that the online experience
was more successful this year than it was in 2010... but if anyone has feedback for making it better in the future, let me know.
In my mind, there are a few themes and outcomes worth mentioning:
1.) Linked data -- in 2010, we briefly discussed linked data but didn't really see how it fit in the VuFind picture. In two years, things have advanced significantly, and now its role seems more significant. I think a point is approaching where we will need
a) a way of representing VuFind records as linked data and b) a way of crawling linked data and formatting it into ingestable records. All of this may become more clear as new standards emerge from the Library of Congress. In the meantime, we really need
to keep this on our radar. I'm still going to be focused on finishing VuFind 1.4 and 2.0 for a few more months, but this may become the next major priority. I'll probably start putting it on the developers call agenda so we can begin focusing our ideas.
2.) VuFind 2.0 release date -- various factors are making me feel like a mid-2013 release for VuFind 2.0 makes the most sense. We need to set a date so that people can start planning to implement the new code, and next summer offers some conferences where
it might be possible to make an official release announcement. In order for this to be successful, the design needs to settle down fairly soon, and I could use more feedback on what works and what doesn't. Hopefully we'll have more people looking at 2.0
code now that the Summit is over. Please don't be shy about sharing your thoughts, whether positive or negative.
3.) Over-engineering vs. under-engineering -- obviously, VuFind 2.0 is a more "software engineered" product than 1.x. Zend Framework 2 forces this, as does the emphasis on extensibility without core code modifications. The Summit reinforced the fact that
we have all types of developers working on the project, from experienced and formally-trained software engineers to librarians who dabble in scripting. Ideally, I want a design that won't make an experienced programmer cringe and also won't scare away a relative
novice through excessive complexity. It's a difficult compromise to make, and one that has to rely on good step-by-step documentation in some cases. Once again, this is a request for feedback; fresh eyes on the code will help reveal problems I've failed
to notice, and more voices in ongoing design discussion will help keep us on the right course. Once the final 2.0 version is released, I will be much more reluctant to make fundamental changes to the code, so now is the time to get it all right!
Thanks again for everyone's participation in the project -- with your help, I think we're on track to make next year's release worth the wait.