Performance on long book texts is poor

[leojonathanoh]

Current

Performance on long book texts is poor

Expectation

Performance on long book texts is not poor

Discussion

Performance on long book texts is poor e.g. >=10000 characters.

Related: #240 #249 #7 #251

As suggested in #249, there needs to be a way to update the Trainer Speech with less changes in the DOM, to ensure Response and Speech performance does not stutter.

A possibility is to break up the book text into chunks of 10000 characters. Each chunk will correspond to a HTMLElement / HTMLTextElement. On book init, the Speech element is updated with these elements. Upon each Response, Trainer Speech feedback is only generated for the current chunk (i.e. zone the Response cursor is in) and only the corresponding HTMLElement updated. In such an implementation, there is only at most 10000 character elements updated in the DOM on each Response event, which should greatly enhance App responsiveness for long book texts.

• One downside is that the Student will not be able to view Speech feedback in peripheral chunks.
  • However, arguably this downside only applies to non-perfection, since in perfection the Speech never shows more than one error (i.e. error at the cursor).