Potential conflict between German and English parser resources when run simultaneously

[GoogleCodeExporter]

What steps will reproduce the problem?
1. Load a German pipeline with the following resources:
prs-ger-cs_1.model";
tagger-ct.model";
lemmatizer.model";

2. Load an English SEPARATE(!) pipeline with the following resources:
prs-eng.model";
tag-eng.model";
lemma-eng.model";

3. Now parse a German sentence with the parser from 1. and inspect the output 
(everything is fine):
1 - Karin - Karin - SB - 2 - fliegt - NE - 
2 - fliegt - fliegen - ROOT - 0 - ROOTnode - VVFIN - 
3 - nach - nach - MO - 2 - fliegt - APPR - 
4 - New - New - PNC - 5 - York - NE - 
5 - York - York - NK - 3 - nach - NE - 
6 - . - _ - PUNC - 2 - fliegt - $. - 

4. Now parse an English sentence with the parser from 2. and inspect the output 
(everything is fine):
1 - This - this - SBJ - 2 - is - DT - 
2 - is - be - ROOT - 0 - ROOTnode - VBZ - 
3 - nice - nice - PRD - 2 - is - JJ - 
4 - and - and - COORD - 3 - nice - CC - 
5 - pretty - pretty - CONJ - 4 - and - RB - 
6 - . - . - P - 2 - is - . - 


5. NEW: Again use the parser from 1. and parse the German sentence (OUTPUT 
CONTAINS ERRORS NOW AND LOOKS STRANGE!!!):
1 - Karin - Ka - ROOT - 0 - ROOTnode - NNP - 
2 - fliegt - flieg - P - 1 - Karin - POS - 
3 - nach - nach - MNR - 1 - Karin - NNP - 
4 - New - New - APPO - 1 - Karin - NNP - 
5 - York - York - APPO - 1 - Karin - NNP - 
6 - . - . - P - 1 - Karin - POS - 

Any suggestions? Any help is appreciated.

Original issue reported on code.google.com by nikoschenk on 6 Mar 2012 at 1:54