[sword-devel] Unicode for greek and hebrew?

Harry Plantinga sword-devel@crosswire.org
Mon, 7 May 2001 16:55:31 -0400


Remapping a module from one specific font to Unicode is easy
in principle but a bit tricky in practice. I have worked on a
prototype perl program to convert from some specific fonts
(e.g. SIL Galatia) to unicode, and I have a mapping file for
several fonts. However, the program isn't really complete
and the mapping file is not yet complete or thoroughly tested.

The problem is that I am using a web browser to display the
results, and current browsers still have incomplete implementations
of display for unicode characters in some of those way-up-high
extended-greek ranges.  Most notably, composed characters
in unicode (made up of a unicode character and another unicode
non-spacing accent character, e.g.) don't display correctly in any
browser as far as I know.

What software were you planning to use for displaying Unicode?
It also may not handle all Greek and Hebrew correctly.

Worse, the unicode spec is reported to be missing a couple of
cantillation points needed for biblical hebrew.  People are
working on that problem even as we speak.

-Harry Plantinga


> -----Original Message-----
> From: owner-sword-devel@crosswire.org
> [mailto:owner-sword-devel@crosswire.org]On Behalf Of Martin Gruner
> Sent: Monday, May 07, 2001 4:10 PM
> To: sword-devel@crosswire.org
> Subject: [sword-devel] 1.52 & a request
>
>
> Gentlemen,
>
> May I ask what still has to be done before sword 1.52 can be released? It
> would be nice if the release would be before the 23th, otherwise
> we will have
> to offer a cvs snapshot for download to be used with BibleTime 1.0. This
> would be possible, but a official release would be nicer. Thank you for
> considering.
>
> It seems to be my fate to be the one who is always asking strange
> questions
> and requesting new things. Sorry.
> After 1.52 there should be a major change in the way modules are created.
> They should no longer be created based on FONTS (like symbol) which use a
> proprietary and fontspecific encoding. They should rather be
> created based on
> official ENCODINGS (e.g. iso8859-7 for greek). This way users can decide
> which font they want to choose. It will even be possible to use an unicode
> font to display every possible language. It is easy to map the
> standardized
> charsets into unicode in the frontends. But that does not work with
> fontspecific encodings.
> Do you understand what I try to say? Imo this is something that
> should really
> happen, because it is just sensible.
>
> I have just looked into my sword-devel archive. There has been a
> discussion
> in Feb 2000 concerning Unicode support in sword. Waves went high,
> but nothing
> happened until now. IMO moving to the correct language specific charset
> encodings would be a first step into this direction, and not a
> line in sword
> code would have to be changed, because we still stick to 1 Byte
> characters.
> Only some of the modules would have to be recoded, which could be
> done almost
> automatically using GNU recode, I guess.
>
> Martin
>