Wrong Pearse code with suffix -e

[Ansa211]

bin/words sancte perfide improbe

The Pearse code on the "e SUFFIX" line is for some reason 01, not 05 as expected.

I assume one has to look around src/words_engine/words_engine-list_package.adb lines 479--485 and 725--750, but I was not able to solve the issue.
(It's also not clear to me why this is not handled through ADDONS.LAT.)

[mk270]

thanks again .. please state the exact reproduction steps, and the expected and observed behaviour.

[Ansa211]

Setup Words to DO_PEARSE_CODES, e.g. by copying the whole WORDS.MDV file like this:

echo "HAVE_STATISTICS_FILE              N
WRITE_STATISTICS_FILE             N
SHOW_DICTIONARY                   N
SHOW_DICTIONARY_LINE              N
SHOW_DICTIONARY_CODES             Y
DO_PEARSE_CODES                   Y
DO_ONLY_INITIAL_WORD              N
FOR_WORD_LIST_CHECK               N
DO_ONLY_FIXES                     N
DO_FIXES_ANYWAY                   N
USE_PREFIXES                      Y
USE_SUFFIXES                      Y
USE_TACKONS                       Y
DO_MEDIEVAL_TRICKS                Y
DO_SYNCOPE                        Y
DO_TWO_WORDS                      Y
INCLUDE_UNKNOWN_CONTEXT           Y
NO_MEANINGS                       N
OMIT_ARCHAIC                      Y
OMIT_MEDIEVAL                     N
OMIT_UNCOMMON                     N
DO_I_FOR_J                        N
DO_U_FOR_V                        N
PAUSE_IN_SCREEN_OUTPUT            N
NO_SCREEN_ACTIVITY                N
UPDATE_LOCAL_DICTIONARY           N
UPDATE_MEANINGS                   N
MINIMIZE_OUTPUT                   N
START_FILE_CHARACTER             '@'
CHANGE_PARAMETERS_CHARACTER      '#'
CHANGE_DEVELOPER_MODES_CHARACTER '!'" > WORD.MDV

bin/words perfide ends with:

01 e                    SUFFIX                  
06 -ly; -ily;  Converting ADJ to ADV
01 perfid.e             ADV    POS                         
02 perfidus, perfida, perfidum  ADJ   [XXXBX]  
03 faithless, treacherous, false, deceitful;

The initial 01 in this part of the output should be 05 (it's a SUFFIX line), but as it comes from the code mentioned above and not from the ADDONS.LAT file, it gets a wrong code.

[ids1024]

What are Pearse codes based on? Googling "pearse code" doesn't help. Are they a standard thing (in computational linguistics, etc.), or just used in words?

"Pearse" must come from somewhere (someone's name, probably)...

[mk270]

I'm guessing Roger Pearse. The classics / linguistics / compsci nexus on this github issues board is occasionally useful :)

[mk270]

Thanks @Ansa211 , I can now reproduce the problem. What is the basis on which you know the correct Pearse codes? Is the system documented somewhere?

[Ansa211]

I don't think it is really documented, but from experimenting with it a little bit, this is what I expect:

01 stem.ending      morphological analysis
02 dictionary forms         [dictionary codes]
03 translation
04 used only for unknown words
05 used for affixes: this line appears before lines 01,02,03, and gives the affix
06 this code has double meaning (which is a bit stupid):
        in conjunction with 05, it gives the English translation of the affix
        without 05, it gives information about a trick used to analyse the word 

I think that for the 06 lines of the second type to appear in the output, one has to have at least one of the following lines in WORD.MOD:

DO_COMPOUNDS                      Y
DO_TRICKS                         Y

and for the 02 lines, DO_DICTIONARY_FORMS Y has to be there.
Then try words abdias and abierant to see the second version of the 06 code.

Running grep '\b0[0-9]\b' * */* */*/* */*/*/* on the directory with Words seems to imply that there are only two places in which codes are output, the files src/words_engine/words_engine-list_package.adb and src/words_engine/words_engine-parse.adb (in the second one lines with code 00 seem to be output; but I haven't seen them in the actual output, and I don't know of any setting which would switch that on).

[mk270]

Please check out commit ac417de which breaks the Pearse Code machinery into its own module

[Ansa211]

I've checked it out; the pull request has been updated to reflect these changes.