Hi Fran,
Thank you for helping us to test LanguageForge! We appreciate the detailed description of what you've had trouble with, and we'll do our best to fix the issues you've encountered. I have a few follow-up questions about each issue you've had, so that we can figure out what is causing the problem and how to fix it.
1) When you try to type dʒəɡɽa, I assume you're using an IPA keyboard. But we actually have two different versions of the IPA keyboard, one that uses Keyman, and one that uses the standard Microsoft keyboard system (called MSKLC, which stands for MicroSoft Keyboard Layout Creator). These keyboards work in different ways, so if LanguageForge has a problem with the Keyman keyboard but we try to test it with the MSKLC keyboard, we might not see the same problem. To type IPA, are you using a Keyman keyboard, or an MSKLC keyboard?
2) Sort order is indeed a problem in LanguageForge right now, and we're working on it. Yes, the goal is to match what FLex does. To help us test whether we've gotten it right, could you give us a list of a few words in Devanagari script from your project, and what order they *should* be sorted in? About 10-12 words should be enough for us to test with. If there are words that computer programs often get wrong, it would be particularly useful to us if you could make sure to include several such words in that list. (For example, in the Thai language, some vowels are written to the left of the consonant but sorting is based on the consonant, and programs often incorrectly sort based on the vowel because it looks like the "first letter". If Devanagari has anything like that, then including several words of that type in the test-word list would be very useful to us.)
3) We don't handle variant forms very well yet, but we're working on fixing that too. One of our thoughts is that LanguageForge is intended to be a simpler counterpart to FLex, not a replacement for FLex, and so some features of FLex will be deliberately omitted. So we're considering changing the LanguageForge import process to only import main entries, and not import variant forms or complex forms at all. We'd like to get some feedback on that idea: do you think that that would be a better approach for collaboration? That is, if variant forms and complex forms just aren't visible in LanguageForge at all, would that be better for you in the way that you would be collaborating with others on the LanguageForge project?
Thank you for your feedback. We hope to use it to make LanguageForge an even more useful tool for your needs.
In Him,
Robin Munn