----------------------------------------------------------------------------

Updated Grammar Rules and the Translation Results

-----------------------------------------------------------------------------

Your program will be producing internal representations of sentences.

The Internal Rep (IR) and updated grammar rules that include calls to semantic routines for the 5 cases of are as follows.

Please note that #s are used to indicate names of semantic routines.

e.g.

::= #xyz# means

In a, call b, call the semantic routine xyz, and then call c.

In the following IRs, italics is used for description of actual values while

Upper Case letters are used for scanner token types of some reserved words.

For Case 1:

::= [CONNECTOR #getEword# #gen#]

#getEword# SUBJECT #gen#

#getEword# #gen#

#gen#

PERIOD

IR to be produced is:

CONNECTOR: english-word

ACTOR: word or english-word

ACTION: word

TENSE: VERB or VERBNEG or VERBPAST or VERBPASTNEG

e.g. for "rika wa naki masen" (there is no connector)

ACTOR: rika

ACTION: naki

TENSE: VERBNEG

For Case2:

::= [CONNECTOR #getEword# #gen#]

#getEword# SUBJECT #gen#

#getEword#

#gen#

PERIOD

IR to be produced is:

CONNECTOR: english-word

ACTOR: word or english-word

DESCRIPTION: word or english-word

TENSE: IS or WAS

e.g. for "rika wa shiawase desu" (there is no connector)

ACTOR: rika

DESCRIPTION: shiawase

TENSE: IS

e.g. for "dakara watashi wa kanashii deshita" (there is a connector)

CONNECTOR: Therefore

ACTOR: I/me

DESCRIPTION: kanashii

TENSE: WAS

For Case 3:

::= [CONNECTOR #getEword# #gen#]

#getEword# SUBJECT #gen#

#getEword# DESTINATION #gen#

#getEword# #gen#

#gen# PERIOD

IR to be produced is:

CONNECTOR: english-word

ACTOR: word or english-word

TO: word or english-word

ACTION: word

TENSE: VERB or VERBNEG or VERBPAST or VERBPASTNEG

e.g. for "rika wa kurasu ni iki masu"

ACTOR: rika

TO: kurasu

ACTION: Iki

TENSE: VERB

For Case 4:

::= [CONNECTOR #getEword# #gen#]

#getEword# SUBJECT #gen#

#getEword# OBJECT #gen#

#getEword# #gen#

#gen#

PERIOD

IR to be produced is:

CONNECTOR: english-word

ACTOR: word or english-word

OBJECT: word or english-word

ACTION: word

TENSE: VERB or VERBNEG or VERBPAST or VERBPASTNEG

e.g. for "rika wa ringo o tabe mashita"

ACTOR: rika

OBJECT: ringo

ACTION: tabe

TENSE: VERBPAST

For Case 5:

::= [CONNECTOR #getEword# #gen#]

#getEword# SUBJECT #gen#

#getEword# OBJECT #gen#

#getEword# DESTINATION #gen#

#getEword# #gen#

#gen#

PERIOD

IR to be produced is:

CONNECTOR: english-word

ACTOR: word or english-word

OBJECT: word or english-word

TO: word or english-word

ACTION: word

TENSE: VERB or VERBNEG or VERBPAST or VERBPASTNEG

e.g. for "shikashi watashi wa okane o kare ni

age masendeshita"

CONNECTOR: However

ACTOR: I/me

OBJECT: okane

TO: he/him

ACTION: age

TENSE: VERBPASTNEG

NOTE:

1) Notice that some parts of the Internal Representations are still in

Japanese. This is because your dictionary is incomplete. Thats fine.

2) We are NOT writing the generator which transforms the above Internal

Representations into actual English sentences.

That would be a separate program.

----------------------------------------------------------------

How to get English Words from the Lexicon/Dictionary:

Tasks for semantic function #getEword#

-----------------------------------------------------------------

For each entry in the Internal Representation of a sentence,

either a Japanese word or its English equivalent must be displayed.

To do this we will use #getEword#.

Whatever #getEword# retrieves will be saved and then be later used by #gen#

As soon as each CONNECTOR, or is found by the Parser

(before we lose the current lexeme), call #getEword#.

Hint: current lexeme saved in saved_lexeme.

#getEword# looks up the Lexicon/Dictionary and saves the English word in saved_E_word..

But if the word is not in the Lexicon/Dictionary, just save the Japanese word saved_E_word.

The exact places in your code where you call #getEword# depend on

how you re-wrote the grammar rules. But it has to be right after

CONNECTOR was matched, was called, or was called.

-------------------------------------------------

How to produce IR with #gen#

-------------------------------------------------

In the grammar rules above,

#gen# marks the places in where

generation of a line of the Internal Representation has to occur.

e.g. ACTOR: rika

The output from #gen# should go to the file translated.txt.

Make sure this file is new each time you use the parser.

In , right after CONNECTOR is seen,

#gen# generates the CONNECTOR: line with #getEword# result of CONNECTOR

In , right after SUBJECT is seen,

#gen# generates the ACTOR: line with #getEword# result of

In , right after ,

#gen# generates the DESCRIPTION: line with #getEword# result of

#gen# generates the TENSE: line with scanner token type of the tense.

Hint: current token_type is found in saved_token.

In , right after OBJECT is seen,

#gen# generates the OBJECT: line with #getEword# result of

In , right after DESTINATION is seen,

#gen# generates the TO: line with #getEword# result of

In , right after is seen,

#gen# generates the ACTION: line with #getEword# result of

In , right after is seen,

#gen# generates the TENSE: line with scanner token type of the tense.

Hint: current token_type is found in saved_token.

The exact places in your code where you call #gen# depend

on how you re-wrote the grammar rules.

Note that #gen# has to be passed what kind of line it is to generate

e.g. gen(description of line).

Given the kind of line and the saved_E_word or saved_token a line can be generated.

How do we write the grammar rule for it??

There has to be 8 cases.