[sword-devel] new/updated modules

Chris Little sword-devel@crosswire.org
Wed, 6 Jun 2001 13:45:47 -0700


> thank you for the information. I believe morphological analysis will
be
> very
> important for sword later. What direction are you moving to concerning
> this?

I agree.  I'd like to see more original language tools & scholarly works
integrated into Sword.

> What about morph searching (e.g. search all optatives)? Is that a
frontend
> issue?

I don't know yet, but I suspect this would mostly be in the library.  We
can do something similar to the strip filter but have it strip
everything except morph tags.  We should do the same for Strong's
numbers.

> The GBF spec has only a small (actually unusable) number of morph
tags.
> *snip*
> Should we try to get this updated (new gbf spec) using your grkmorph
> definition rather than putting proprietary tags into gbf?

I took the "to be expanded" to mean "define your own".  The three tags
it includes are legal tags for the definition I used.  This wouldn't be
the first time I've done invalid GBF (lines ended in CR rather than
CR-LF, >254 characters per line, & maybe UTF-8 encoding being some
examples).  If we want Michael to change GBF, we may also want to beg
him to make it XML compliant.

There are 3 different Morphological tagging systems that I know of, and
I don't see any reason why we shouldn't allow all three and do
translations between them like we do with different text markups. The
three systems are:

Friberg's - used in his Analytical Greek New Testament (I haven't seen
this system so I don't know how similar it is to the others)
Packard's - used on CCAT for the morphological GNT, LXX, & BHS (I also
have the LXXM in this tagging system)
Robinson's simplified version of Packard's system - used in OLB texts
including BYZ and TR; this is what LXXM uses.

I asked the compiler of the LXXM why he used Robinson's rather than
Packard's, since the former omits some information encoded by the
latter, and he replied:

	The difference is primarily in the stem and generally has little
	impact on the semantics and rather has to do more with how the
stem is 
	written in view of the ending it's connected to. In Bible study
I'd be 
	more concerned with the meaning, the semantics, rather than
form. The 
	fact the Dr. Robinson thought it unnecessary to include the
extra stem 
	information in his modules is sufficient for me to suggest
leaving it 
	out also.

He went on to cite compatibility with the other OLB modules and
readability as reasons to use the simplified tags.  I'd rather not lose
information, so I believe I will use the original Packard tags in the
next version of LXXM.

Does anyone think I should not do this or believe we should standardize
on a single tagging system?


> And what about
> hebrew?
(http://www.shamash.org/tanach/tanach/text/bhs.info/bhsmorph.txt?)

Hebrew tagging would just be another system to add to the mix.  Getting
a distributable, tagged text will be another matter.  The only one I
know of online is the CCAT text and I believe they have some unpleasant
redistribution restrictions.

--Chris