Background

OfficiumDivinum came to be when, having started an attempt to generate LaTeX booklets from divinumofficium and given up trying to work out how it worked (in the end I used a scraper to get the content, although with difficulty), I (John) discovered that Endre was running his own divinumofficium clone, with the source edited to produce better output. Both of us wanted something better, and I had a working start for a rendering model, and Endre a head for unintelligible perl (and rubrics), as well as web code. We decided to put together a complete re-write, as much to learn how to do it as for the end product.

Where divinumofficium made everything bespoke and mostly unextendable, we made wrapped everything in objects and implemented output with rendering methods, any number of which could be created for differing output formats. This is very flexible, and should make building applications around it a doddle, but it comes at the price of conceptual complexity and cpu cycles needed to turn a DSL into objects, solve precedences, render a bunch of objects to some output format, and then serve that output format to a user. We work around this with caching, but there are certainly more efficient (if perhaps less readable) ways to build this.

We also determined that OfficiumDivinum should be heavily documented, and have a fully-functional API.