An alternative CW learning method
Basically, I want to experiment with an approach slightly different to the Koch method. Instead of blasting 20 words or so at the student at once, I want to subject the student to single words at full speed and proper spacing, having the student type the full word only at the end of each listen, and then make the words longer as the accuracy improves.
My idea is that this, unlike conventional Koch, might be better for learning head copy, as it encourages the student to copy and remember longer and longer stretches of code as opposed to having them become CW stenographers.
- Make and enter a Python 3 virtual environment.
- Install the required dependencies:
pip install -r requirements.txt - Run the thing like so:
./scatterbopper.py koch_level word_length test_length
Basically how many unique symbols will be in the test. Increment when you want to introduce additional symbols. Currently uses the symbol sequence described in the 1936 Koch study, which isn't considered the best sequence nowadays, but it was the first one that came to mind. Will add other sequences later.
The length of the word you will have to listen to before being given the chance to type your answer. Set to 1 in order to learn individual letters first; set to 2 to learn to head copy letters in pairs, and so on, working your way up.
The amount of words you will be tested on before being presented with your score.
The following features are proposed:
- Make the learning sequence changeable/configurable.
- Incremental mode: continuous testing with the word length auto incrementing or decrementing based on how well you're doing. New letters to be introduced in cleartext at first, until the user has gotten them right a couple of times.
- Testing for non-letter symbols and prosigns.
- Callsigns mode: Variable-length words consisting of callsigns as they would appear in the wild.
- Optional real-life confounders: QRM, QRN, QSB.
- An Android port would be neat, but that'd be a bit out of my wheelhouse.