[sword-devel] Greek texts

Paul Gear sword-devel@crosswire.org
Tue, 18 Jul 2000 22:05:25 +1000

Chris Little wrote:
> With the exception of our locked NA27/UBS4 B-Greek text, which was I believe
> produced by Troy, I did all the Greek texts using the files from Broman.
> There are, I've been made aware, problems in the character mapping and
> probably the transliteration as well (things like Chis and Xis swapped).

Grrr!  Don't get me started!  :-)  Actually you have gotten me started -
see below.

> So I really wouldn't mind revisiting the whole Greek text subject and
> redoing them all.

Well, neither would i.  I want to make them much more readable (with
capitals, punctuation, etc.) and include morph codes, though. 
Interested in working on a database?  I've already collected some texts
and made a start on the data model.

> I'd also like to get away from using Symbol as a font, even
> though it's pre-loaded on every Windows box.

Easy on Windows - just provide a different font.  Hard on Linux - what
other font are you going to use?  How are you going to ensure that it's
present on everyone's Linux box?  Are you going to force frontend
developers to use FreeType and a particular TTF, or just let them work
it out themselves?  Are you going to make everyone use xfsft?  What
about other platforms?

> I'm certainly open to suggestions of what new texts and character mappings
> should be used to create new versions of these.

For theta, psi, chi, and upsilon, use the scheme adhered to by the
B-Greek discussion list
<http://metalab.unc.edu/bgreek/transliteration.txt>.  They do not
distinguish the final sigma.

There are commercial fonts that use the B-Greek mapping:
- Bwgrkl (from BibleWorks, i think)
- Linguists Software's Graeca/GraecaII (from http://www.logos.com)
- Galaxie Software's Greek <http://www.bible.org/galaxie>.
The latter two map the final sigma to double quotes ("), while Bwgrkl
uses lowercase v.

The free fonts that are available around the place have some rather
silly mappings:
- SIL Galatia maps H to rough breathing and J to eta (i understand why,
but i think lookalike should override soundalike in this case, since
rough breathing isn't a letter)
- OLBGrk maps Q and Y the other way around
- SPIonic swaps X and C (God only knows why)

I've recently found a couple of (seemingly free) Greek fonts that use
the B-Greek mapping:
- EGreek
<http://www-user.uni-bremen.de/~wie/Egerton/egerton-greek.html>.  No
license seems to be attached to this.
- GreekFP <http://www.paradigm-sw.com/gfp/font.htm>.

The latter is a better quality font and uses the same final sigma as
"Graeca" and "Greek", but the vendor probably needs to be contacted to
ensure that we would be able to redistribute the font.  It says it is
available for non-commercial use only (which presumably we qualify for),
but there isn't any indication as to whether redistribution is permitted
or anything like that.

I think B-Greek encoding is the way to go - make people who want to use
something different convert with a font map.  This gives the best chance
of readable output if no Greek font is available.

Final sigma is a hard one - you could either encode it as a double
quote, or not distinguish it and let the frontend programs deal with the
problem by mapping an S at the end of a word to the appropriate
character for their font (or maybe do nothing).

"He must become greater; i must become less." - John 3:30