Steps forward, steps back, on entry rendering
: Martin Holmes
Minutes: 240
Made a good start on the entry rendering, I thought, by creating an HTML template system for the output tables which can be easily completed by templates matching custom processing instructions. However, after successfully doing the first table of three expected for verb entries, I came up against some real complexity in the second, and it transpires, after some discussion with ML, that a fair amount of actual specific knowledge will be required to know how to structure tables for individual entry cases. This knowledge should not be encoded in processing rules; instead, it should be part of the encoding, only acted on by the processing. So we’ll need to think again and devise some more structures for encoding what amounts to domain knowledge, and then we can use it to build the tables. Meanwhile, we’re a bit pushed for time and will probably have to leave out these complex bits for next week’s discussion.