[Biblical-languages] Text processing of Greek in a large batc h of files

Peter Kirk peterkirk at qaya.org
Wed Sep 10 07:21:57 EDT 2003


On 10/09/2003 04:26, biblang at ntresources.com wrote:

> > There is a document TECkit Language.doc.pdf which comes with the
> > downloadable TECkit package.
>
> Thanks. I spent a couple hours studying that last night and drafting a 
> tentative .map file for Galilee. I have a few questions that are not 
> clear to me. If anyone is familiar with this and could offer me some 
> pointers I'd appreciate it. I've searched the documentation and sample 
> files and not found any explanation (at least that I can understand; 
> perhaps these things are clear to programmers, etc.).
>
I'll try to answer this, but I am also a learner, not an expert - I am 
currently working on a map for Hebrew. I will also copy this to Jonathan 
Kew, who is the expert, in the hope that he might correct any mistakes I 
make.

> 1. Should the header field, "RHSFlags" be set to "(ExpectsNFC)" if I'm 
> converting to the pre-composed characters in the Greek Extended range? 
> (The sample Galatia file has ...NFD.) I'm not clear as to what this 
> does; perhaps it only impacts the Unicode > byte direction?
>
I think it is only for the Unicode > byte direction, and means that this 
reverse mapping will work whatever form is presented to it. If your RHS 
mappings are all preomposed, you should use GeneratesNFC and ExpectsNFC. 
But it might be easier for you to set up all the mappings for NFD and 
get TECkit to precompose the characters for you with its conversion 
flags or a special normalisation pass.

> 2. Is the following syntax correct for mapping multiple legacy 
> characters to single Unicode values?
>
> 'a' 'v'  <>  U+1F70  ; alpha acute
> 'a' 'j' '/'  <>  U+1F80  ; alpha smooth iotasubscript
>
I think you should have just 'av' and 'aj/' on the left, but the two 
syntaxes may be synonymous.

> (I suspect that such substations could be done with classes, but since 
> I'm the one who has to write this table and then debug and support it, 
> I figured doing it manually, one at a time, would be easier for me to 
> figure out; the class idea sounds efficient, but I'm not clear on the 
> intricacies of it. I also chose the "U+..." format followed by a 
> comment tag for RHS values to save typing so many long names as 
> "greek_small_letter_alpha_with_oxia".)
>
> 3. Assuming that #2 is a legal way to make such multiple 
> substitutions, would I assume correctly that these values should 
> precede single character conversions? That is, change all the multiple 
> sequences first, then change any single characters not already 
> converted. Or does the "more specific"/"longer rule" priority make 
> that unnecessary?
>
Good question. I would expect the system to automatically process longer 
strings first, as that is what many users are used to from the CC 
program, but I can't find it stated.

> 4. And one final perplexity: what function does the vertical line 
> indicate, as in the second half of the Byte_Unicode pass in the sample 
> Galatia map? e.g.,
>
> '|b'  <>  greek_beta_symbol  ; curly beta
>
> I've searched the documentation in vain to figure that one out. There 
> is a note on p. 9 of the language doc that refers to "handling 
> characters that are represented in SIL Galatia as sequences using the 
> '|' modifier code, but map to single Unicode characters." But it never 
> tells me how this works. And the rule I've just cited certainly isn't 
> a "sequence" of characters; it's a single code position...? (This is 
> one thing that makes me wonder if I've understood the syntax in #2 
> above correctly.)
>
If I remember correctly, '|b' etc are character sequence which could be 
used with SIL Galatia on the Mac to display alternate glyphs, not 
accessible from the PC with that font. You would need to see the 
documentation for SIL Galatia (not Galatia SIL) for that.

> Many thanks for any help on this.
>
> Rod
>
> = = = = = = = = = = = = = = = = = = = = = = = = =
> Rodney J. Decker, Th.D., Assoc. Professor/NT
> Baptist Bible Seminary, Clarks Summit, PA, USA
> URL: www.NTResources.com 
> PURL: purl.oclc.org/NT_Resources/
> Email: <rdecker> at <NTResources.com>
> = = = = = = = = = = = = = = = = = = = = = = = = =
>  
>
>------------------------------------------------------------------------
>
>_______________________________________________
>Biblical-languages mailing list
>Biblical-languages at lists.ibiblio.org
>http://lists.ibiblio.org/mailman/listinfo/biblical-languages
>  
>


-- 
Peter Kirk
peter at qaya.org (personal)
peterkirk at qaya.org (work)
http://www.qaya.org/




More information about the Biblical-languages mailing list