[sword-devel] .conf lang tags
Thu, 7 Nov 2002 16:48:44 +0100
what you describe is certainly possible, but not satisfying. Besides the fact
that I'm not sure how much work parsing the unicode range would be (icu?) for
each single key, there are languages which operate in the same unicode range
but the user might want to set different fonts for them.
A solution without requiring extra markup for each key might be to have some
kind of entries in the conf file indicating which key range(s) is _not_ in
the module's main language and if so, which language. That would be the
easiest way coming to my mind.
But I understand that you hesitate given the work that needs to be put into
this, and the unclear (at the moment) benefit. It's not the most serious
issue right now, though I believe it might become somewhat important in
Am Mittwoch, 6. November 2002 01:55 schrieb Troy A. Griffitts:
> If your frontend really wants to support swapping fonts based on the
> content, you can derive the context based on the character ranges used.
> For example, if you examine the characters used in a key for an entry
> and decide they are all in the Thai unicode pages, then you know you
> need a font that has the Thai pages populated. I'm not quite sure
> adding support at a lower level of granularity than "all the code pages
> used in this module" is ideal. It would mean adding much additional
> markup that can be derived just the same by looking at the character
> itself. Does this make sense?
> Martin Gruner wrote:
> >> From first thought, I would suggest keeping the single language tag
> >>that represents the primary module language, and add addtional AuxLang
> >>tags which represent other languages that are contained within the
> >>module. This will allow a frontend to select a unicode font with all
> >>pertinent pages populated.
> > Hey Troy,
> > that might not always be possible. Imagine a chinese-korean dictionary,
> > and a user having a special font for chinese (only) and another one for
> > korean (only). How should he work that out?
> > Providing the possiblity to use a unique font for each language is our
> > aim with BibleTime, because this might be required in certain situations.
> > Our previous paradigm was having one Latin-1 font, and one font (like
> > Code2000) for all non-Latin-1 modules. That worked to a certain point,
> > but is definitely not the optimal solution, because even Code2000 does
> > not support all languages (like Chinese) correctly, or render each
> > language as nice as possible.
> > Therefore we need a way to determine the language of each single key of a
> > lexicon, because it may have keys in different languages. Or another
> > solution which would avoid the problem I described above. I am not so
> > much asking this for myself or the western Latin-1 world, but for the
> > users who come from other areas of the world.
> > Martin