Coffee|Code : Dan Scott

Well, I'm finally back from Atlanta and the Evergreen AcqFest. I'll apologize right off the top for not providing more blog updates over the course of the weekend, but the requirements and design discussions were pretty intense so I didn't want to risk continuously missing subtle but important details and not being able to participate intelligently by live-blogging the event. After each full day of work, we "unwound" with a serious meal--which, after some socializing, usually involved slipping back into kicking more design and implementation ideas, problems, and potential solutions around. By the time I got back to the hotel room, I was either completely wiped out, or itching to commit something to the group document / or play with some code. So I hope you understand (all three of you that are reading this!).

On top of everything else, it was a bit of a gruelling trip back. In order to save a few hundred dollars and make a greener transportation choice for the final leg of my journey, I took the bus back to Sudbury. Hello, five-hour layover in Toronto and a packed six-hour bus ride (thanks to Hwy 401 construction) back home! It was quite a relief to get back and see the family.

Anyways, here is a mini-summary of what we accomplished:

• Agreement on some realistic time frames for acquisitions and serials development: call the stages one philosopher, two philosopher, and three philosophers
• After kicking the left-of-field idea of using a calendar server to handle serials schedules, coverage information, predictions, and claiming events for a day or so, and clearly invoking the concern of at least one library blogger, Mike had a brilliant idea for how to represent all of this natively in PostgreSQL. He's going to take a few days to work through a proof-of-concept to ensure that it's as solid as it sounds, so I won't give away the details just yet...
• Agreement on the requirements for basic item-at-a-time acquisition workflow support (to be implemented first) and more advanced acquisitions support (batch orders via MARC record import, integrated vendor discovery API support, EDI support)
• Agreement on adding internationalization support to OpenSRF. Right now OpenSRF (the messaging infrastructure on which Evergreen depends) knows nothing about locales. We've been able to use URL tricks to support translation of the catalog interfaces thus far, but Mike worked through the changes that will be required to pass locale as a property of each session. This will enable the service being invoked to "do the right thing" if locale is of a concern to whatever output it returns.
• _Almost_ agreement on how to add internationalization support to Evergreen. We worked through a number of different scenarios for supporting translation of dynamic strings (library names, for example) that reside in the database, from a single table that holds all of the translated strings, to an i18n schema that holds tables that parallel any table in another schema that holds translatable content. We settled on the latter schema. I say "almost agreement" because until something gets committed to code, I have a feeling that this is still subject to change a little bit
• Exposure to some parts of Evergreen that many of us hadn't seen before -- in particular, the reporting interface that Evergreen provides is extremely powerful and well-designed. It even supports basic line and bar charts for adding punch to your presentations.
• Art introduced us to OFBiz and OpenTaps via an online training video, and later on Art and Ed successfully played around with the Java OpenSRF client via BeanShell. My takeaway lesson about OFBiz if I ever need to customize something built on it: it's all in controller.xml!

There was a lot more than that that we covered, but for now I have to get some shut-eye. For any skeptics out there, the actual acquisitions and serials workflows employed at our constituent libraries were used as testbed scenarios for the discussions about Evergreen's serials and acquisitions requirements and design. I'm feeling good about the work we accomplished, I think we found some elegant solutions for some of the age-old problems in these areas, and I think we have a common understanding of the path forward.

Bill Denton, you are a genius!

So, my colleague Bill Denton recently landed a job as the Web Librarian at York University. While chatting idly with him, err, strengthening my professional network, I related a tale of mock woe and bitterness about how York had turned me down for a job at their Scott Library when I was attending library school.

A few days later, Bill sent me this image envisioning a horrible alternate reality where I got that job, and I laughed, and laughed, and laughed a little bit more. So I must share it with you, and praise Bill, and wonder just how things would have turned out had I spent my first year of library school providing first-line reference services to patrons, rather than manually converting SGML to HTML. Hmm... work requiring social skills, or obscure technical skills? I think we all know how that turned out.

So I'm taking off tomorrow for Atlanta to spend four days deeply immersed in discussing, designing, planning, and implementing Evergreen's acquisitions and serials support. At least that's the plan. In our spare time (heh), we're going to tackle the internationalization infrastructure as well. The spirit of the event is modelled loosely after the Access conference hackfests, and has therefore been dubbed Acqfest. Unlike hackfest, however, where the journey itself is usually the goal, with Acqfest there's a much stronger emphasis on actually getting things done. I may, err, acquire a slight southern accent after a few days, but I mostly hope to increase my understanding of Evergreen while kicking in some design suggestions, code, and documentation here and there.

Laurentian is covering my travel--at this point in our evaluation of the future of our systems, it's in the library's interests to give me an opportunity to stare deep into the heart of Evergreen--and my local arrangements are being covered by BC Public Libraries, Georgia PINES, and Equinox Software. I'm contributing my time and, uh, expertise. All round, I think the whole community is going to benefit from the Evergreen Acqfest. Assuming I have a few minutes, I'll try to post some updates on our progress over the next few days.

So, yesterday afternoon Mike Rylander from the Evergreen (a.k.a Open-ILS) project pushed out the first release candidate of Evergreen 1.2.0. Hurrah! If you tried installing Evergreen before, but got hung up on some of the build, install, or configuration steps, I think you'll find this release a lot easier to deal with. For example, there's one less configuration file to deal with now -- bootstrap.conf is a thing of the past.

I'm happy to point out that I've updated my Gentoo-based VMWare image of Evergreen as well: Evergreen 1.2.0-rc1 (479M). Along with that, I've updated my instructions for installing Evergreen on Gentoo to reflect the newer, simpler install process.