Part of offering help to language learners is presenting flexible techniques; some people prefer flashcards, some prefer fill-in-the-blank questions, some like to read a lot of examples. Some like to see grammar spelled out, some like to grab colloquial usage and eventually come to understand the rules. Some people don't really know their preference.
One idea for the simplest data model I can imagine that allows freedom in learning is:
Lojban Sentence Submission (by anyone): lojban, natlang, and comments on one page
Use a blog-like web-page format of sentence case-studies with associated comments to provide a web-accessible learning database for lojbanistas (anyone interested in lojban).
Scenario 1
* Lojbanists who already know their stuff can put submit lojban sentences with either/both linguistic, thorough glosses and natural language (possibly multilingual) renditions.
* Beginners can search by
keyword, vocab word, grammar, etc. They find the sentence, they ask a question in the comments, and something like what happens on this mailing list gets recorded in an easily visible webpage form
Scenario 2
* Beginners can submit a lojban sentence and an attempt at natlang translation, or vice-versa, or even just a natlang sentence they want to know how to say in lojban.
* Lojbanists fill it in, and use the comments to discuss, like this mailing list, any tricky grammar points etc.
Scenario 3
(maybe a bit more fun)
* Same as scenario 1 or 2, except the lojban sentence leaves a _____ blank to be filled in. A lojbanist might use this to write a 'textbook problem' or to overcome jolban writers block, and a beginner can use this technique to answer questions, or test the expressive limits of lojban.
Several days ago I mailed a data model for a language-lesson format. This solution seems to be a bit more
tailored to lojban, because it seems the bulk of lojban conversation takes this form naturally (a text excerpt, then a collaborative linguistic dissection with interspersed grammar questions). The two structures could overlap (sentences as a searchable content type are used in both).
Any thoughts/suggestions?
I can't say I prefer any one scenario over the others - they all look good to me.
However, the important point is that different people learn in different ways.
For me, the reading of texts is the most important.
The problem with this is that there's a bunch of texts on the wiki, with no
differentiation between completed well-written ones that have been vetted
by several experienced lojbanists, and ones that are poorly written or incomplete.
Obviously a beginner should start off with a 'good' text and leave the ones that
need vetting till later. I understand that once there was a plan to have 'officially
approved' texts, but that is probably impractical at this time.
So I suggest that the website includes some recommended texts, with the
recommendation based on some reasonable criteria.
totus