Re: LYNX-DEV LYMessages_en.h translation

[T.E.Dickey]

> 
> to follow up on what T.E.Dickey said: 
>  
> > > One potential danger is that what doesn't get distributed doesn't  
> > > get used, and falls out of sync quietly so that you don't fix it  
> > > and when you do need it, it doesn't work.  
> > yes - but if we do the gettext (or whatever) integration properly, those 
> > extra files are just data, like 
> >  
> >     00001 Hello world 
> >     00002 Do you wish to quit 
> >  
> > (and once assigned, we _don't_ change the numbers). 
>  
> What is your opinion of 
the scheme above is how I remember gettext's organized (Jim Spath did say
he's done a conversion for that - he'd know offhand, I'd have to check). 
We'd continue to use the symbols that are in the header file.  (I'm not too
fond of those integers, but it's easy to convert for the sake of maintenance).

>       A: opaque irrevokable integers as above and vector retrieval 
>       (code not mnemonic) 
>  
>       B: sorta-English message names as at present and  
>       hash the access of the translations.  (program biased 
>       to U.S. English base, OtherLanguages 2nd class). 
>  
>       C: mnemonic+integer message names.   (long names) 
>       (integer extract plus vector access for retrieve) 
>  
> ?? 
>  
> Al 
>   


-- 
Thomas E. Dickey
address@hidden
http://www.clark.net/pub/dickey