Halfway through this one week hackfest, it is time to tell you more about what is happening at BibLibre this week!
No less than 35 collaborators have come together to make Koha better. And this time people have come frome Germany, Croatia, Italy, Switzerland Spain and even the United-States to participate. People who work in libraries that use Koha have also come from all over France to bring us the functional approach we need to validate that everything is working smoothly.
We organized ourselves in 5 groups.
The biggest one is the one composed of people who work in libraries. It includes librarians, but also ILS administrators or system engineers who work with Koha. Since they are less constantly looking at code than most of others, they bring us a new, very functional, way of looking at things! They have worked on translating the interface and manual, and have tested a lot of patches.
Another group is working towards making Koha database system(RDBS)-independent. At the moment, Koha can only work with MySQL, but by cleaning up all the parts that are specific to this RDBS, and by making a few other parts better, we will then be able to use PostgreSQL for instance. This means more flexibility and a better separation of code layers.
Some of you already know about it: another indexing engine has been installed for a few libraries that use Koha: Solr. This has been a lot of work, but the result was definitely worth it: better search functionalities, more efficient, and most importantly extremely flexible, with the ability to define indexes and facets in a few clicks! Another big project is now taking place: bringing the possibility to the community version of Koha to use Solr instead of Zebra, the current indexing engine. This is a long term project, and nothing will be visible after the hackfest, but we are setting important foundations for the months to come.
A workshop is also going on around ergonomy: there are many small things in the Koha admin interface that deserve to be made better: this group concentrates on these problems. Until now, we have standardized terms used, we spot inconsistencies in the interface and normalize practices. Our goal: more consistency and a smoother interface. And has a bonus: a more beautiful interface! We have added classes and unique ids which will allow to target specifically a page or all the pages of a module with css: it will soon be possible, for instance, to use a color code to identify each module!
The last group is working on persistency: integrating Plack will allow for a massive improvement in performance. But for this integration to work, some code needs to be rewritten in order to be compatible with Plack. The first benchmarks run on what has been done so far indicate a 5 fold performance improvement!