All this in
gtsvn/ped/doc/testing/

Candidate list:
egrep ' \?' corpus_vasta_pos_divvun_multi | cut -d"\"" -f2
(the words which are not recognised by the analyser)
this gives a list of MISSPELLEDWORDs


For each MISSPELLEDWORD, find the divvun suggestions in
divvun-misspellings_090505.txt 

The best, dynamic solution, is:
For each MISSPELLEDWORD, generate the suggestions via Tomi's command:

cat MISSPELLEDWORD | ./private/trunk/polderland/bin/spellSamiNort -u8 -0 -m /Users/tomi/langtech/gt/sme/polderland/SamiNortLexWithCatalanRez   > MISSPELLEDWORD.out

cat MISSPELLEDWORD.out lookup -flags mb ~/gtsvn/gt/sme/bin/ped-sme.fst | lookup2cg | spelleroutput2cginput.pl > misspelledword.cohort

and then place it back in corpus_vasta_pos_divvun_multi (or wherever it came from)

"<MISSPELLEDWORD>"
	"MISSPELLEDWORD" '\?'	

CHANGES TO:

"<MISSPELLEDWORD>"
	misspelledword.cohort

----------
eventually cheating (i.e. not via Xserve):
use the file divvun-misspelling...

content for dunje:

scan throu file divvun-misspelling to
' MISSPELLEDWORD...
drop mext line

Prompt: Check returns 0[0:0]/1, @0:5 for 'dunje'
Getting suggestions for dunje... <=== wordform
Suggestions:
247	Tonje  
247	Tonje-
245	dune
(...)
240	dolje
240	dolje-
240	donbe
Prompt: Check returns 0[0:0]/1, @0:5 for 'duina'


grep with initial munber and etc

take output and | lookup -flags mb ~/gtsvn/gt/sme/bin/ped-sme.fst | lookup2cg | spelleroutput2cginput.pl | see

... put it in  corpus_vasta_pos_divvun_multi 
where it belongs (under MISSPELLEDWORD