mhulden / foma

Automatically exported from code.google.com/p/foma
117 stars 90 forks source link

An FSA that behaves strange #52

Open GoogleCodeExporter opened 9 years ago

GoogleCodeExporter commented 9 years ago
What steps will reproduce the problem?
1st case:
1. load stack strange.fst
2. view net  (*)
3. write prolog strange.prolog
4. read prolog strange.prolog
5. view net  (**)

2nd case:
1. Compile hun_cg2.rlx with fomacg
2. load stack strange.fst
3. define Sentence
4. read prolog R_1_43.prolog
5. define Rule
6. read regex Sentence .o. Rule;  (*)
7. read regex Sentence;
8. random-upper  (**)
9. pop
10. read regex Rule;
11. apply down (**)

What is the expected output? What do you see instead?
1st case:
The net in (*) is linear; the net in (**) is star-shaped.

2nd case:
The composition in step 6 (*) does nothing -- Sentence remains unchanged. 
However, if you apply down Rule on Sentence's random-upper (one word, since 
it's linear), Rule correctly #X#'s out the non-pronoun readings of "<az>".

What version of the product are you using? On what operating system?
0.9.17alpha
Latest fomacg.

Please provide any additional information below.

Original issue reported on code.google.com by nemesk...@gmail.com on 30 Jul 2013 at 2:45

Attachments:

GoogleCodeExporter commented 9 years ago
The fomacg part is so that you can reproduce the problem. The point is: there 
is sentence where apply down and .o. produces different results, even though 
they shouldn't.

Original comment by nemesk...@gmail.com on 30 Jul 2013 at 2:46

GoogleCodeExporter commented 9 years ago
OK, disregard the fomacg part, it was due to the fact that the space wasn't 
part of the wordform symbol. However, the 1st case still stands.

Original comment by nemesk...@gmail.com on 31 Jul 2013 at 8:41