Changes from previous versions of J_Inflector
0.3.0 released June 9, 2000:
- Added the ability to choose multiple items from the menu when
using the -m switch. Multiple menu entries can now be listed
in space delimited form to select more than one inflection.
One side effect is that inflection output can be reordered this
way, because they appear in the order selected.
(As suggested by Nicolas Simonet)
- Menu items now number from 1 instead of 0, with 0 being
reserved to cancel the command.
- Added the capability for the -m switch to take an argument
which causes an output similar to picking off the menu without
displaying the menu. This is a shortcut if you already know
the menu numbers. Input is in bar delimited form (e.g.
1|2|3). No spaces are allowed unless the argument is in
quotes.
- Enhanced the rule screening mechanism to allow for
auxiliaries. For instance, for verbs that conjugate
with suru or nasaru, these endings would be
recognized for instance if the user enters "yakusuru"
or "ohanasinasaru". The set of rules applying to
those auxiliaries is invoked. Suru was already covered in
previous versions, but the mechanism has been generalized.
Allowable auxiliaries are now listed in Rules_VClass.dat.
- Created a mechanism in the rule elimination process
to allow for exceptions to irregular exceptions for
verbs conjugating according to an auxiliary. For instance,
suru has a "regular" conjugation, but there is also a
set of irregular exceptions patterned on "suru" that
applies to verbs like "aisuru". Suppose that there
was such a verb "Xsuru" that conjugated mostly like "aisuru"
but had one or two exceptions to rules that were already
patched. Now a further patch, specifically for "Xsuru"
can be applied, and will screen out the other irregular
rule that was already overriding the normal rule for suru.
That is, any auxiliary irregular rule is eliminated in favor
of any irregular rule SPECIFIC to a given verb.
- Added the capability to create "blocking rules" which can
block conjugation to a particular tense for a specific
verb or adjective.
- Enhanced the verb ruleset to complete (so far as I know)
the "aisuru"-like conjugations. Users should be aware that
I don't have a good handle on what tenses are affected,
but at this point the potential and all forms derived from
-nai have been given exceptions.
- Enhanced the verb ruleset to include the remaining irregular
honorific verbs (gozaru, nasaru, etc.).
- Added rules for additional potential and causitive passive
inflections for regular itidan and godan verbs.
(As suggested by Nicolas Simonet)
- Removed the "colloquial polite" politeness level in favor
of a "colloquial" designation that can be applied to any
rule to denote that the result of that rule may not be
strictly grammatical or formally correct speech.
- Removed the "humble" and "honorific" politeness levels
to their own "attitude" structure. These qualities
denote a different axis of the paradigm, and should be
independent from level of formality. Attitude denotes the
speaker's feeling _about_ whom they are speaking, and
politeness or formality level denotes feeling toward the
person _to_ whom they are speaking. Conjugations can now
be marked humble or honorific, and still show a separate
politeness level. Attitude is sticky, so that once conjugated
into humble or honorific, a word retains this quality as it
is further conjugated.
- Altered the verb dictionary file to record an infinitive-form
kanji as well as a stem-form used for conjugation (for instance,
as in the case of ossyaru). Additionally the dictionary now
records the inherent attitude of a verb so that conjugations of
inherently honorific or humble verbs can be so marked.
- Added a word history to cache previously conjugated words.
- Added storage of the results of the last conjugation.
- Added a new command cw (for "current word"). By itself,
it will print the current word as a reminder. With the
switch -h, it will set the current word by letting the
user select from the last 10 conjugated words. With the
switch -r, it will allow the user to select a current
word from the last set of returned results (note that if
the last conjugation only had one result, that result is
set as the current word automatically.
- Corrected the documentation and error messages on the form
of roomaji being used. It is a modified form of Nihon
roomaji, not of Kunrei. For instance, the syllable that
is "ti" + nigori is rendered as "di" not as "zi".
Roomaji "zi" is reserved for "si" + nigori. The
documentation now shows a table of syllabic equivalents.
0.2.0 released June 6, 2000:
- Made last word entered be the current word, if not replaced
by a result. This lets the user conjugate into multiple
inflections without having to reenter the verb being worked
on. (Planned improvement: the addition of a command to
select and capture any one of multiple returned results
as the current word.)
- Changed the argument entry for inflection specification to the
switch -t to prevent confusion when using switch -f and to
make the command syntax more explicit.
- Fixed output to file so that adjective conjugations can also
be exported.
- Brought the adjective conjugator to functional par
with the verb conjugator, including reading from a rule
file (Rules_Adj.dat) and a dictionary file (Words_Adj.dat),
and output to EUC with kanji/kana. Created limited versions
of these two data files.
- Added support for kanji and kana EUC output of inflectional
forms. Inflectional forms now appear in the file created
when switch -f is specified. Inflections appear in kanji
(when available from the dictionary) plus kana ending,
and a full kana-only reading. Accuracy for kanji
processing of irregular conjugations may be spotty.
0.1.4 released June 5, 2000:
- Externalized verb classifications to a Rules_VClass.dat
file to make the classification engine accessible to
change by the user. A list of potential suru verbs that
might have irregularities similar to "aisuru" is included
below the data lines, and can be cut and pasted by the
user into the data area if needed.
- Created facility for external dictionary files to keep
verb or adjective entries. This will enable kanji/kana
processing, differentiation between homophones & explicit
classification capabilities.
- Created a small test dictionary file for verbs
(Words_Verbs.dat).
- Added a check of the dictionary file on each non-forced
conjugation, which then presents all available alternatives
to the user (as well as the option to use the built-in
classifier instead). With a sufficient word file, this check
will allow the user to choose from between homophones which
may conjugate differently. Furthermore, it will allow for
kanji/kana output of the resulting conjugation.
Added a switch (-f) to allow export to an EUC coded text file.
Ultimately this will include the kanji (if the word is
pulled from the word file) and kana inflected
forms. For the time being, where kanji and kana inflections
would be, the verb's dictionary form kanji/kana is simply
echoed out.
0.1.3 released June 4, 2000:
- Added support for automatic recognition of godan verbs
ending in -iru and -eru.
- Fixed designation of "mairu" as the result of the humble
inflection of "iku" and "kuru" so it is labeled godan.
- Added support for the irregulars aru and kureru.
Note that for kureru meaning "grow dark", user must
force conjugate to itidan (-i switch).
- Added support for recognition of one-kanji Sino-Japanese
-suru verbs which are slightly different in some inflections.
The screen and ruleset is not complete, so the only one currently
handled is "aisuru".
- Added switch -rs to force the "regular" suru conjugation in case
any of the irregular -suru verbs have regular homophones.
- Added conjugation of the copula when the verb specified is
"da" or "desu". Rules are not complete, but a few basic
tenses are available.
- Added the adjective conjugator and enabled switch -a.
Rules are not complete, but a few basic tenses are
available.
0.1.1 released June 1, 2000:
- Fixed recognition of words with standalone 'n'.
- Got rid of the WinSIOUX save question when app is closed.
- Put in some of the rules to keep certain conjugations from
building on others, i.e. the taberareraremasu problem.
- Made a major improvement to the irregular exception mechanism.
- Completed adding exceptions for "iku" to the data file.