Ok, a new version is up now.
changelog wrote:2/26/10
- Wordlists can now be loaded from or saved to a text or CSV file.
- Partial wordlists can now be loaded from existing dictionary files.
- Language name is now stored in savefile (set it on the last tab)
- It is now possible to opt out of phonotactical restrictions on the phonotactics
page.
- Sanity check: Attempting to generate more than 1000 clusters on the phonotactics
page now results in an error message (if you need more than about 500, you're probably
better off just opting out of phonotactics restrictions anyway).
- The dialog for choosing sequences of phonemes has been optimized for displaying
larger number numbers of sequences (YMMV based on your screen size).
- Deleting phonemes and suprasegmentals that are referenced in many places will now result
in all references being properly removed rather than a crash.
NB: The last line is not
quite true - I noticed, after I had finished compiling in Windows and packaging everything up, that I'd forgotten to fix a file. All it means is that for now, after you delete phonemes, you should go re-save the phonotactics before doing anything else (in particular, regenning word phonologies, though loading wordlists will probably also crash it). I'd upload the really really fixed version instead, but since I don't have access to Windows right now, it'd just be in linux, and I want to avoid too much version confusion, if possible.
Anyway, new features:
1. You can turn off phonotactics. Note that this will cause all of your words to be phonology-less (for now). Also note that after turning off the phonotactics on the phonotactics tab, you will have to actually click "save phonotactics" for it to take effect.
2. You can save to text files. CD will write each word entry to each line of the file (though if you have newlines in your descriptions, those will be preserved). The formula entered in the formula box determines what is written - /w means the spelling of the word, /t means the type, /s means the subtype, /p means the phonology (the text that's displayed on the word tab) and /d means the description. Any other character will be written literally. Using the formula "/w","/p","/t","/s","/d" (yes, with all the quotes and commas) will produce a CSV file that can be read by excell (though you may have to change the extension to .csv).
3. You can read wordlists from text files. The formula you enter for that is essentially the same as the above (ETA: except for /p - phonology can't be directly read from .txt files), but characters other than the forward-slash-escaped characters and parentheses will be interpreted as a
regular expression. Non-punctuation characters will still be interpreted literally, though, so that's not really relevant unless you actually want to make use of regexp functionality. Note, however, that quantifiers don't seem to work with the forward-slash-terms (e.g. "/s?" will match the subtype followed by a question mark, not an optional subtype). There might be other quirks to it too, especially in Windows. CD will spawn a pop-up telling you how many lines of the file were actually matched, and if it turns out to be 0 or something, you can cancel.
4. You can import wordlists from other .cdic files. The phonologies of these words will be ignored, and they will be reinterpreted based on their spelling (this is because there's no way to guarantee that the phonologies correspond, at the moment). You should be able to select consecutive words simultaneously by pressing shift and either scrolling with arrow keys, or clicking, and non-consecutive words with ctrl-click.
5. You can name your language on the last tab.
Let me know how all of the new GUI stuff looks.
(Sorry)
I'll do that tomorrow, and then hopefully figure out what I want to do with the parser over the weekend maybe.
