ConlangDictionary 0.3 - now phonology parsing is faster

[faiuwle]

Sure, just add both romanizations to your phoneme, and it will figure it out no matter how you spell it, as long as the romanizations don't overlap with the romanizations for other phonemes.

[Qwynegold]

You mean what you posted in this thread? I'd need the whole file to really test things, though I have been slacking for a bit lately, so I don't need it right now. I do have a sample db file that came with LP, but I'd rather look at what actual conlangers are doing with it.

No, I mailed you some last week.

[faiuwle]

Odd, I don't see it. Send it again?

[jal]

Odd, I don't see it. Send it again?

Such exchanges could better be addressed via a PM.


JAL

[Qwynegold]

OK, sent it again. (Sorry jal but I can't be bothered with PM for such a small thing to say.)

[jal]

OK, sent it again. (Sorry jal but I can't be bothered with PM for such a small thing to say.)

Sometimes Q, you are a lazy idiot. There's a "PM" button in the interface that will give you the exact same screen as the "Reply" button, except that you PM the person instead of everyone.


JAL

[Torco]

Boo Hoo

[jal]

Boo Hoo

I've got kids. Sometimes I get in this teaching manners mode.


JAL

[Torco]

I was a teenager recently; sometimes I get into this "dislike manner-teaching" mode.

[faiuwle]

Ok, new version available from the original download links. It's just the Lexique import stuff that's new. Click on "Edit Paradigm Settings", and you can add paradigms and put what fields they correspond to. The order is important - the paradigms listed first will be expected first in the entry. If you assign multiple paradigms to the same field, it will use the value specified by the last one that isn't empty, and all fields except for the word itself are optional (they'll just be blank).

For example, by default the paradigms are
\lx = word
\ps = type
\ge = defiition
\de = definition

so stuff will parsed like
\lx fari (Word)
\ps verb (Type)
\ge eat (Def)
\de

or
\lx fari (W)
\ps verb (T)
\de eat (D)

or
\lx fari (W)
\ge eat
\de eat, transitive (D)

etc. Whitespace and unrecognized paradigms shouldn't matter, as long as everything it does recognize is in order.

I did get a crash before when launching the Edit Paradigm Settings dialog, but after putting in debug messages I haven't been able to reproduce it. If it happens, let me know what you did. Also, when I loaded a 13000-line .db file, it actually took a few seconds to load, so don't be alarmed if you have a huge file and it seems to freeze momentarily.

I'm going to start 0.3 the weekend after next, so let me know about any bugs before then.

[Shinali Sishi]

Is it bad that I broke it already? I was trying to edit the phonology because no matter the length of the word or the use of IPA symbols/regular letters it keeps showing /???/ for the phonology, and it crashed when I clicked "add syllable." I set up the phonemes and the phonotactics before I went there. And I forgot to save so I lost everything I did with those. -_- What am I oing wrong or is it just my old dinosaur computer that hates phonologies?

[faiuwle]

Odd, that hasn't crashed for me, yet. Do you have phonotactics turned on? Did you save the phonotactics on the phonotactics tab? If so, it should auto-analyze the words based on the spellings when you import.

ETA: Oh, I see. It happens only when you try to edit a word that doesn't have a phonology because you haven't saved the phonotactics yet. You should save the phonotactics before importing, but I'll fix it so it at least doesn't crash.

[faiuwle]

Right, crash should be fixed now, though you still won't be able to change the phonologies from /???/ without actually adding at least some phonemes. This actually came from the last update, when I added the phoneme-moving buttons and forgot to check for the case when there weren't any syllables.

[Shinali Sishi]

I clicked save but I may have not had the touchpad register that. Thanks! I really like this program so far.

Edit: is there a reason the autophonology likes to ignore my word final codas? Like gem it lists as /ge/ and chanchan as /tSan.tSa/

[Qwynegold]

Haven't tried the newest newest version yet so I don't know if these things have been fixed already, but...

1. Whenever I add a blank line, after I quit the program and then start again, it will have added several blank lines there. So for example

Code: Select all

foot; leg [inanimate, oblong]

See: hat'ka-k'ōngya

becomes

Code: Select all

foot; leg [inanimate, oblong]











See: hat'ka-k'ōngya

And the extra blank lines seem to be cumulative.

2. When you start the program, it automatically says the name of which conlang's name comes first in alphabetical order in the load and save as fields. I think this is really dangerous because you could accidentally save over another conlang's dictionary. I think that when you load a dictionary, the same name should automatically be displayed in the save as field.

3. Is there a way to delete a type or subtype? I had "time" as a subtype of adverbs, but I want to change it to "temporal".

[Qwynegold]

Oh, thought of some more (I hope I'm not piling too much on you, these are merely suggestions):

4. It would be nice if there was an ignore characters function. For example in one dictionary I have spaces and dashes that don't correspond to any phoneme (for example I have entries such as "kyappo-langp'o" and "negefak'p'ö k'ūllössen"), so I have to first write the entry as one word and then change the spelling without updating the phonology. If it was possible to have the program ignore certain characters it would be more user-friendly. (The space is quite problematic, but I think one's not really supposed to use it in phonemic presentation, so you might as well ignore this problem, but... the above word, negefak'p'ö k'ūllössen, I've managed to get the program to display as /ne.ɡe.fɑkʰ.pʰø. kʰuːl.løs.sen/ by making the space character a suprasegmental, but the dot after the syllable pʰø is not supposed to be there.)

5. (I might be the only person who would have use of this, so this issue is not so important.) It would be nice if one could have [] instead of //. There's three different ways one could implement this:

A. Just an option to change all // to [], or all [] to //.
B. To be able to specify for each entry if you want to use // or [].
C. To be able to have both phonemic and phonetic representation of the same word.

A I guess would be the easiest thing to program and C the most difficult. I myself would only need A right now.

[faiuwle]

Edit: is there a reason the autophonology likes to ignore my word final codas? Like gem it lists as /ge/ and chanchan as /tSan.tSa/

I don't know; I can't reproduce it. What are your phonotactics? Is the spelling being correctly read from the file?

1. Whenever I add a blank line, after I quit the program and then start again, it will have added several blank lines there.

I can't reproduce this either. You're not still using 0.1, are you? Try downloading the latest version, and if it still does it, let me see your savefile.

2. When you start the program, it automatically says the name of which conlang's name comes first in alphabetical order in the load and save as fields. I think this is really dangerous because you could accidentally save over another conlang's dictionary. I think that when you load a dictionary, the same name should automatically be displayed in the save as field.

Yeah, that's probably a good idea. I'll add a confirmation if you try to overwrite a different file, too.

3. Is there a way to delete a type or subtype? I had "time" as a subtype of adverbs, but I want to change it to "temporal".

Just go through and find all of the words listed as "time" and change them to "temporal" - the list is just a list of all the strings currently being used as types and subtypes. Type and subtype are going to get overhauled when I do features, at which point it should probably be easier to manage them globally.

4. It would be nice if there was an ignore characters function. For example in one dictionary I have spaces and dashes that don't correspond to any phoneme (for example I have entries such as "kyappo-langp'o" and "negefak'p'ö k'ūllössen"), so I have to first write the entry as one word and then change the spelling without updating the phonology. If it was possible to have the program ignore certain characters it would be more user-friendly.

Yeah, I can do that.

(I might be the only person who would have use of this, so this issue is not so important.) It would be nice if one could have [] instead of //.

What are you trying to do? Are you wanting to use phones instead of phonemes for everything, or do you just want to specify phones when they differ significantly from the phonemes? Or something else? I could
1) make a global setting like you suggested
2) make a phonetic field for you to fill in manually if you wanted, or even
3) do something where you could specify a set of allophones and then set the phonetic field via the interface
but I'm not sure I want to set up all the interface it would need to be able to specify allophonic rules and then follow them automatically. It's not that that would be hard, but it's a lot of interface and overhead and it's not really part of my overall design. It would kind of depend if you were going to want something more detailed here in the future (because then it would make sense to at least lay a ground structure with the allophones and the extra field) or if you just want a cosmetic change option.

[Qwynegold]

1. Whenever I add a blank line, after I quit the program and then start again, it will have added several blank lines there.

I can't reproduce this either. You're not still using 0.1, are you? Try downloading the latest version, and if it still does it, let me see your savefile.

Nope, it was 0.2. I'll try version 0.3...

2. When you start the program, it automatically says the name of which conlang's name comes first in alphabetical order in the load and save as fields. I think this is really dangerous because you could accidentally save over another conlang's dictionary. I think that when you load a dictionary, the same name should automatically be displayed in the save as field.

Yeah, that's probably a good idea. I'll add a confirmation if you try to overwrite a different file, too.

Great!

3. Is there a way to delete a type or subtype? I had "time" as a subtype of adverbs, but I want to change it to "temporal".

Just go through and find all of the words listed as "time" and change them to "temporal" - the list is just a list of all the strings currently being used as types and subtypes. Type and subtype are going to get overhauled when I do features, at which point it should probably be easier to manage them globally.

Oh, then I'll wait for that. Oh yeah, this reminds me. It would be nice if the part of speech would be shown on the list of words so that one can quickly find the type of word one is searching for, instead of having to click every entry to see what PoS it is...

[

4. It would be nice if there was an ignore characters function. For example in one dictionary I have spaces and dashes that don't correspond to any phoneme (for example I have entries such as "kyappo-langp'o" and "negefak'p'ö k'ūllössen"), so I have to first write the entry as one word and then change the spelling without updating the phonology. If it was possible to have the program ignore certain characters it would be more user-friendly.

Yeah, I can do that.

Yay!

(I might be the only person who would have use of this, so this issue is not so important.) It would be nice if one could have [] instead of //.

What are you trying to do? Are you wanting to use phones instead of phonemes for everything, or do you just want to specify phones when they differ significantly from the phonemes? Or something else? I could
1) make a global setting like you suggested
2) make a phonetic field for you to fill in manually if you wanted, or even
3) do something where you could specify a set of allophones and then set the phonetic field via the interface
but I'm not sure I want to set up all the interface it would need to be able to specify allophonic rules and then follow them automatically. It's not that that would be hard, but it's a lot of interface and overhead and it's not really part of my overall design. It would kind of depend if you were going to want something more detailed here in the future (because then it would make sense to at least lay a ground structure with the allophones and the extra field) or if you just want a cosmetic change option.

In this particular lexicon I want to have phonetic representation instead of phonemic. If it's not too much trouble I would really appreciate nr. 1.

[faiuwle]

Nope, it was 0.2. I'll try version 0.3...

Not out yet. There were different 0.2s depending on the date. I asked about 0.1 because the save and load stuff was fundamentally different and much more error-prone.

Oh yeah, this reminds me. It would be nice if the part of speech would be shown on the list of words so that one can quickly find the type of word one is searching for, instead of having to click every entry to see what PoS it is...

Yeah, that'd be easy to do.

By the way, is the new phonology-editing stuff working well?

[Shinali Sishi]

I solved the word-final coda problem (actually by remembering a glitch in sounds - a space after the word will make it not recognize the coda as such and it just ignores the syllable! I edited the cdic file and that's happy now.)

Other than thinking the already suggested ignoring spaces/hyphens/parentheses would be just swell, It is working perfectly now!

I like how I can add suprasegmentals and call them anything and have them do most anything (like lyitatata contracts to lyitta so I told it that /t/ can have a geminate, even if it's not a real geminate; also /e:/ and /o:/ romanize weird so I could just add that too).

You know, I never really came up with phonotactics for Quanafi before this program, and itis going to make my life a lot easier on making new words.

[faiuwle]

I solved the word-final coda problem (actually by remembering a glitch in sounds - a space after the word will make it not recognize the coda as such and it just ignores the syllable! I edited the cdic file and that's happy now.)

Other than thinking the already suggested ignoring spaces/hyphens/parentheses would be just swell, It is working perfectly now!

Ahh, whitespace. Yeah, I'm going to make spaces be ignored by default.

I'm glad you like it.

[tezcatlip0ca]

About Italian, I meant <c> for both /k/ and /tS/, depending on its environment...

[faiuwle]

Yeah, sorry, I forgot about your other post. People have mentioned that before, and I'll probably do it eventually, but I don't know if I'll wind up doing it this week.

[Qwynegold]

Nope, it was 0.2. I'll try version 0.3...

Not out yet. There were different 0.2s depending on the date. I asked about 0.1 because the save and load stuff was fundamentally different and much more error-prone.

Oh, so it shouldn't be happening in the version I'm using? I'll send to file to you...

Oh yeah, this reminds me. It would be nice if the part of speech would be shown on the list of words so that one can quickly find the type of word one is searching for, instead of having to click every entry to see what PoS it is...

Yeah, that'd be easy to do.

Thanks!

By the way, is the new phonology-editing stuff working well?

Do you mean the part where you add phonemes? I haven't experienced anything that's wrong with it, except this one thing I can't figure out the meaning of:

When you have the "show parsing order" checked and write stuff in the spellings field, then if you check "show alphabetic order", all the spellings fields are empty, so you have to write all the spellings twice, once for the parsing order and once for the alphabetic order.

[faiuwle]

Ok, but you are using the latest latest version now, right? Sometimes things get fixed by accident, and I don't actually keep the older versions around, so I can't debug things that go wrong with them.

Do you mean the part where you add phonemes?

No, actually - the place where you manually edit the phonologies of the words, since you were talking about turning off the phonotactics.

When you have the "show parsing order" checked and write stuff in the spellings field, then if you check "show alphabetic order", all the spellings fields are empty, so you have to write all the spellings twice, once for the parsing order and once for the alphabetic order.

That's weird, and I can't reproduce that either.

The parsing order is the order the phonemes are considered in when auto-determining the phonemes in words. The alphabet order is just the order that the words are sorted in in the list.