I didn't know there was one ;-)
but there is more than one way to skin a cat.
Which is more directly related to the Stenotype. This must have been fiendishly complex to make mechanically, but the 'programming' is mainly a couple of look-up tables. By my count this handles more combinations than Velotype, while being slightly less easy to learn.
Another table for the vowel keys, and you're done. All an orthographic keyboard has to to os provide for the chording of common letter sequences in a single stroke. All you have to do is learn the chords for the missing letters and you know how to type anything. There are also a few special chords for common syllables that don't fall easily into the key sequences, but you can learn those as you go along. No arrangement is perfect, but it doesn't take much to represent 90% of monosyllabic words and a fair few disyllabic ones in a single stroke.
For example, if you code the ten most common vowel pair onto the four steno vowel keys, that covers over about 90% of vowel pairs and only once in about every hundred words would you need to stroke (for example) suitable as su.it.able.
Good enough? If you were really, really fast you would have to do that two or three times a minute, which would be annoying. By then you would be thinking of either another key or two to handle more cases, or dictionary briefs for your most common exceptions. With an orthographic system the only complete words and phrases that need to be in the dictionary are the ones that you put there yourself.
Because there is no system dictionary, and therefore no multistroke dictionary entries you need to indicate word boundaries manually, either as Dothan does, with a space key that can be stroked simultaneously with the letters; or as Velotype does, by assuming that every multi-letter chord is a monosyllabic word unless they are joined by a No Space key. The latter cuts the usage of the most commonly-used key in half.
For full keyboard emulation and more I think we one Meta key to provide layers. When it is pressed the lefthand (or righthand) keys are check and the result used as a index to swap in alternative lookup tables. These would either remain in force as long as the Meta key was held down, or the Meta key used on its own would restore the default tables.
So the decoding sequence is:
Check special keys and change table if needed
Concatenate the strings indexed by the main keys.
Is it a brief? If so, replace it.
Apply modifications suggested by the special keys (capitalisation, space, punctuation)
Output.