Patrik Nyman wrote:
>> I'd feel very inclined to store the text twice, once broken
>> down by pages, one by structural divisions.
>> That way the hideous task of isolating the interesting
>> elements in between two <pb> is done once, not on every retrieval.
> But wouldn't breaking down the text by pages require an XPath
> expression of equal hideousness to the one required for querying
> by pages?
yes, but doing it once and making a static copy
would make writing the queries easier. Just a
thought, I am sufficiently glad I don't have to
solve it myself...
If you search the archives of TEI-L, good and hideous
XPath has been posted for this "locate a <pb/>, now
extract everything up to the next <pb/>" problem.
Either from Martin Holmes or Michael Beddow, as I recall.
If its not on the TEI wiki at
whoever finds it first should put it there...
Information Manager, Oxford University Computing Services
13 Banbury Road, Oxford OX2 6NN. Phone +44 1865 283431