[sword-devel] Getting stuff done (Re: External links)

Ben Morgan benpmorgan at gmail.com
Tue Nov 25 13:56:53 MST 2008


On Wed, Nov 26, 2008 at 6:46 AM, Karl Kleinpaste <karl at kleinpaste.org>wrote:

> All you do is to convince me, again and again, not to generate OSIS.
>
> I absolutely do not care if it wouldn't pass muster "within an osisRef
> attribute."  It passes muster NOW in ThML TODAY, because what I generate
> gets USED by people who wanted it YESTERDAY.
>
> ThML, to me, is about Getting Stuff Done, and using tools that
> facilitate Getting Stuff Done, as opposed to wanking over whether I've
> accommodated the Syntax Purity Demigods.
>
I'd have to say that I don't like ThML. Too much like html. ThML makes it
much harder for the frontend developer (me) who wants to process the text a
bit more...

BPBible 0.4 uses an XML parser to parse text for indexes - well, for ThML
and OSIS modules anyway (when can we get rid of GBF?). By having osisRef's
which are consistent much processing time can be saved. With ThML there are
two different forms of the passage tag, and they can really have whatever
the writer wants in them. This then has to be parsed, etc. I also noticed
that ThML modules are also much more likely to be invalid XML - which most
parsers can't really handle. And then if they aren't XML, the fallback
mechanism doesn't allow searching on xrefs and strong's numbers.

The sync tag is another thing I don't like about ThML. Rather than wrapping
round the text, it occurs at the end. And rather than having 1 tag with
(say) 2 strong's numbers and 2 morph numbers, it will have 4 empty tags at
the end. Much harder to process if you want the user to be able to search on
combinations of strong's and morph.

Also, using ThML makes it harder to guess what the meaning of the markup -
rather than saying we have a catchWord, for example, we may put it in
italics. And if we have included text (like in the KJV) and put that in
italics as well, we presumably can't tell which is which. This type of thing
makes advanced searches like "find the word "..." where it is added by the
translator" impossible. OSIS lets you do extra formatting more easily. For
example, BPBible displays poetry more like a print Bible would. Rather than
having complicated markup in the module, all it needs is a <lg> tag and then
<l> tags within it.

Another example - I've been wondering about using the ESV quote data to put
different people's quotes in different colours. This would be quite easy to
do as the ESV clearly marks what is a quote. There is still the question of
what to do for nested quotes, but the actual inserting the colours would be
easy.

However, I can see that ThML may be better for user editable content, as
they just want to see bold text. But as we know, in this case the user would
be wrong ;)

What really needs to be done, is to finalize how we are going to use
osisRefs.
I'd like to see modules being able to contain some mapping from osis ids to
display names. I don't show the user Gen.2.2 in my application (or
/Gen/2/2), why should I show them Ant.1.8.2 or (/Ant/1/8/2)?

God Bless,
Ben
-------------------------------------------------------------------------------------------
Multitudes, multitudes,
   in the valley of decision!
For the day of the LORD is near
   in the valley of decision.

Giôên 3:14 (ESV)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.crosswire.org/pipermail/sword-devel/attachments/20081126/709430f4/attachment.html 


More information about the sword-devel mailing list