[sword-devel] OSIS recommendations to SIL/JAARS
Kahunapule Michael Johnson
Kahunapule at mpj.cx
Tue Jul 24 17:48:07 MST 2007
DM Smith wrote:
> It is good to hear from you again.
Of course, I now feel really stupid, but I'm glad you pointed out the
error of my ways before I went too far down that path. :-)
> Couple of comments:
> <p> does not have a milestoned form.
Arghh. That pretty much messes up the whole idea. I should have noticed
> These do cross verse boundaries
> and the semantics of <p> don't allow for it to simply be ended and
> restarted on verse boundaries. So I see several options: 1) don't
> use <p> at all,
This isn't a viable option unless a good milestonable replacement exists.
> 2) simulate <p> with something like <lb type="x-end-
> paragraph"/> and <lb type="x-start-paragraph"/>
That is kind of ugly... but it could work. Actually, you don't really
need both of the above. You can revert to the rather early HTML form of
<p> with no </p>, which works fine. :-) However, for consistency with
being able to change to a milestoned form of the p element, both should
> or 3) submit a change
> request against the OSIS spec
I think I would rather wait until we had more of a consensus on what we
wanted before doing that.
> or 4) always mark paragraphs between
Bible translators believe that they have the freedom to place paragraph
boundaries anywhere, not just at verse boundaries. They are right.
> I suggest 2 and 3, as 2 will allow a migration as the intent
> is clear. Having paragraphs start and end between verses is allowable
> in all cases.
> You refer to XMLT. Is this a typo? Should it be XSLT?
> With regard to quotes: Quotes frequently cross verse boundaries. This
> means with verses as containers that they need to be in the
> milestoned form: <q sID="xxx"/> and <q eID="xxx"/>. In your sample
> text you have </q marker="'"> This is improper.
OK, so q elements that indicate start and end of quotes should all be
milestoned, and all have marker attributes. Only the q elements that
indicate Words of Jesus within a verse should be in container form. I
just used the container form in the example because the quotes fit in
the verse that time, although they often don't.
> When using milestoned elements it is easy to produce valid but poor
> OSIS with respect to overlapping elements. When considering just the
> milestoned sID/eID pairs, they should not overlap as in:
> <div sID="xxx"/>....<q sID="aaa"/>.....<div eID="xxx"/>....<q
I thought the whole point of milestoned elements was to allow them to
Technically you could make verses overlap each other, too... but I can't
think of any good reason to do that. I'm not sure why you might overlap
div and q like your example in real text, unless the div and q started
or ended at the same point, and someone chose to write them out in
random order. A really good OSIS reader wouldn't care. Putting those
sorts of things on the same stack in the writer implementation should
take care of that bit of aesthetics.
More information about the sword-devel