[sword-devel] ACDCref, a modified ACDC module: hot-linking verse references
Chris Little
chrislit at crosswire.org
Mon Sep 20 09:48:15 MST 2010
On 9/20/2010 8:08 AM, Peter von Kaehne wrote:
>> Von: Karl Kleinpaste<karl at kleinpaste.org>
>
>
>> We're all about Bible software; I think having modules that always link
>> Bible references is, by now and in our world, a base-level necessity.
>
> I think the same would apply to practically all the GenBook modules and commentary modules.
>
> I have found in CPAN a set of Perl modules which appear quite capable of finding references in large volumes of text - and can be configured to do the same on non English text. I have not yet tried them out, but will so shortly.
I have a faint recollection of trying that module and having no success,
sadly.
Having written a regexp to churn through the Catholic Encyclopedia,
looking for references and tagging them as such (an approx. 20 line
regexp!), I can say that it is an exceptionally difficult task to guess
at all of the ways that references will be marked, even within a
relatively homogeneous text. I don't think it's possible to generalize
that kind of tagger to work with any text in a language.
I've worked with an in-house tool developed by another Bible software
developer that was supposed to detect and interpret all references (and
it only cared about English). It worked so unreliably that it had to be
modified to let me tag refs myself so that it only needed to do the
interpretation part.
--Chris
More information about the sword-devel
mailing list