ram

Yes, I had to found a work around the attribute separator. My suggestion about FreeLing was mainly for the possibility to display the token and its attributes as nodes, instead of plain text. But it seems it implies a lot of fixes and that is something that is going to be fix in version 4. Is that the case, is there any draft for the XML output?

Thanks, sorry for the mistake

Thanks for your response! cwb-describe-corpus -s works perfectly. For show cd I still get incomplete information: ===Context Descriptor======================================= left context: 25 characters right context: 25 characters corpus position: shown target anchors: not shown Positional Attributes: <none> Structural Attributes: <none> Aligned Corpora: <none> ============================================================

About the DTD, I will use the same SGML structure but XML compliant. For more ideas about the schema, FreeLing output formats could be an useful resource.

Corpus info output

XML output mode for CQP

Consistent output formatting

SGML invalid structure