[developers] RMRS characterization differences

Sergio Roa sergior at coli.uni-sb.de
Fri Jul 13 23:33:19 CEST 2007


Hi Christopher,

On Friday 13 July 2007, Christopher Rupp wrote:
> You raised the question of whether the character offsets in PET RMRS
> results were being calculated correctly. [...]
> The context is slightly different,  in that I am processing files of
> marked up XML text and feed in the actual character positions in the
> file at the lexical level, so  the numbers are true for the file and
> not the string input.

I just also want to confirm that by applying a raw text input to cheap
without  the option  -tok=fsr, I  have  also found  that the  spanning
ranges are different.

However, I turned back to use a previous svn revision of cheap, namely
the 300, and the character spans don't differ anymore.

Cheers,

Sergio.






More information about the developers mailing list