[sword-devel] Poetry Encoding
dmsmith at crosswire.org
Mon Mar 30 07:33:00 MST 2009
David (Mailing List Addy) wrote:
> On Friday 27 March 2009 08:39:34 am Chris Little wrote:
>> <l> may only occur within <lg>
> So then, are all <lg>'s orthographic or poetic lines? Or at least
> supposed to be only orthographic or poetic lines? I ask so I can
> determine some sort of logic for interpreting <lg>'s and if they are
> poetry/orthographic automatically and formatting accordingly
I don't know that one can conclude that <lg> is solely used for those
two purposes. OSIS has two ways of creating a new line in a flow of
text: <l>...</l> and <lb/>. In earlier editions of OSIS, both could
occur only in <lg>. Now, <lb/> can occur outside of it.
When doing a programmatic conversion of text into OSIS, there is some
level of analysis that needs to be done to map source markup to target
markup. Such a map is best when it can take context into consideration.
But invariably, a map will produce valid, but inappropriate markup.
I find that source material sometimes (frequently???) doesn't care about
correctness of markup but only resultant appearance. E.g. <br/> and
used to break lines and introduce spacing. This might indicate
poetry, but it might also indicate something else. (E.g. While not
pertaining to line groups, I frequently see <p/>, the empty paragraph,
being used to separate paragraphs or add a blank line.)
While <lg> may have been designed for poetry and such, it can be used
for any purpose. E.g. a FAQ with Q and A being separate lines and each
pair being a line group.
There are no predefined types for <lg>, but it would be appropriate to
distinguish usage on it using x-type (e.g. x-poetry, x-orth, x-faq) but
I'm pretty sure that frontends won't care about x-type attributes as
they are non-standard.
More information about the sword-devel