I noticed that after I had failed to conceptualize the words, what my mind was doing when reading the raw source text from the input field, was it was referencing some of the text up front, and then performing a re-trace, after new information in subsequent sections and sentences.
This seems like a pretty important component, and highly dependent on the writer's style, and tone, which cannot be anticipated programmatically, thus, there doesn't seem to be a silver bullet rule for reading speed, universal to all passages of text.
I always thought that speed reading was non-linear with a kind of gestalt mechanic at work. This seems more like speed listening which is a 1D stream.
With this system you can use the different forms of mental processing of the mind.
Here the code of veolocity calculation equal to the speed-memory
count word is (0 to 14)
_n=velocity;
if(countWord>=3) _n = _n + _n * 0.2; if(countWord>=6 && countWord<9) _n = _n + _n * 0.4; if(countWord>=9 && countWord<12) _n = _n + _n * 0.6; if(countWord>=12 && countWord<14) _n = _n + _n * 0.8;
velocity=_n;
Thanks!