[sword-devel] Verse parsing

[Chris Little]

As Martin said, using '.' as an indicator of an ordinal value is 
EXTREMELY common cross-linguistically.

On the other hand, I don't see why OSIS parsing needs to occur in the 
same parser.  OSIS references are completely controlled and don't 
require the complexity of our general parser.  The occurrence of OSIS 
references is also controlled: they only occur within osisRef/osisID 
attributes in specific elements of OSIS documents.  Besides that, 
without looking at the code, I imagine our parser does not currently 
parse the work portion of an osisID/osisRef, which is completely 
inappropriate for the verse parser.  So why combine them?

--Chris


Troy A. Griffitts wrote:

> Joachim and Martin,
>     I spent some time trying to track down the German verse parsing error
> today and have some information for you.
>
>     To handle the "1. Book" syntax, the parser decides if it has a book
> name yet.  It is does not, it strips the '.' and continues parsing.
> This leaves the book name as "1 Book".  The German locale does not have
> entries that resolve for this.  Maybe you have an argument that the
> parser shouldn't strip the '.', but we are doing it now to conform to
> OSIS, which reserves the '.' for a BOOK, CHAPTER, VERSE separator (e.g.
> osisRef syntax for 1 John 1:1 is '1Jn.1.1').  So, for now, could we just
> get the German locale updated with entries for books without the '.'?
> The user will still be able to enter '1. Johan', but the locale must be
> able to match '1 Johan'.  Let me know if this wasn't clear to you and I
> can explain further.
>
>     Thanks for all your great work!  Looking forward to a release soon!
>
>         -Troy.