[CWB] copying alignments

Stefan Evert stefanML at collocations.de
Wed Dec 16 18:59:02 CET 2015


> On 16 Dec 2015, at 17:50, Ruprecht von Waldenfels <ruprecht.waldenfels at gmx.net> wrote:
> 
> Just to clarify: So that means once the alignment is computed and stored in the alx files, it becomes independent from the structural attribute that was used to define it?

Yes, the encoded alignment attributes refer to corpus positions directly rather than to the structural attribute providing the alignment grid.  That's why it's safer to re-encode alignments in case the token positions have changed.

(Side remark: cwb-align-encode and cwb-align-decode also use corpus positions.  Only the aligner software and cwb-align-import reference a structural grid.  If your alignment comes from an external software, it's definitely safer and better to re-import it for each corpus.)

This is something that's likely to change in CWB4, by the way.

Best,
Stefan


More information about the CWB mailing list