Questions / Comments about storing constructs during IT/IGC

[ggurdin]

1. If I go through IT or IGC, then erase the message I was writing, should those construct uses also be erased?
2. Tokens come from IGC span / IT continuances and tokenization on message send. These two categories are likely to have different offsets, so it's hard to say if the tokens are 'the same.' But it's likely that the tokens of selected choices will be stored twice. What's a good way to mitigate this? I tested with comparing offsets, but the offsets are changed when choices are added to the text.

[wcjord]

1. The learner still did the thinking needed to make those choices so saving the records of it makes sense.
2. After IT, I think we go ahead and assume that all tokens were awarded with construct uses during that flow.
   For IGC, the post-send system calculation of construct uses decides which type to award based on its overlap with an accepted match. You can use that to decide which tokens to exclude so they're not doubled up. So post-send, don't send any that the current system marks as 'ga.' These should have already been awarded in the interactions. Make sense?

[ggurdin]

Yes, that makes sense!