Files
gf-core/lib/resource-0.6/shallow/lexicon/notes.txt
aarne 2ff53b2ffd Started experimenting with shallow resource API.
Started experimenting with shallow resource API.
Added PossessPrep to Structural.
2004-02-20 16:28:42 +00:00

32 lines
987 B
Plaintext

Project: to massage an ispell word list into a GF lexicon. AR 20/2/2004
Used MkLex on linux.words (45424 word forms)
The first run gave 6471 words, which was too much for GF.
With the maxsize, got this down to 1587.
The grammar now compiles in 25sec, and parses as follows:
"every commander assumes of an abbot"
610 msec, 6 ws
"every accusing commander assumes of a coarse abbot"
840 msec, 8 ws
"every accusing commander assumes of a coarse abbot of all
cognitive colonizers"
820 msec, 12 ws
"every accusing commander of a comical careful colt assumes of
a coarse abbot of all cognitive colonizers"
3670 msec, 17 ws (nesting inside)
"every accusing commander assumes of a coarse abbot of all cognitive
colonizers of a colt of a careful carpenter of every carriage"
TIMEOUT!
Note: it is essential to import with the -cflexer flag!
Time spent: 58 min, includes writing MkLex, running tests,
and writing these notes.