Hello again, Rick;
I just left a comment on Lawrence's page that bears mentioning here. He said that 'octonions are really a system of quaternions' (7 of them) which relates to a statement I made in the paper "In Defense of Octonions" with Ray Munroe. I wrote to Lawrence that while octonions could be represented that way, they do have to be resolved in an orderly fashion, and it's not the same as saying O is really H x 7.
In a paper I'm working on now; I suggest that working with octonion algebra is similar to assembling a watch. "Every layer or sub-assembly must mesh correctly, and then the layers must fit together in the correct relationship, for the watch to function. The same metaphor aptly describes what is required to do multiplication and division with octonions, as you must perform seven ordered groups of three operations in sequence."
Is this an apt characterization? My guess is that Lawrence's approach would treat the component quaternions in the same way that Physics folks normally treat grad, div, and curl - as independent or fundamental quantities, where in reality (or as you demonstrate) they are structured components of the quaternions. I suggested his statement might be made true if octonions are treated as an ordered or nested system of seven quaternions. Is this essentially correct? Do we also need an extra scalar value, for the Real component?
Enlighten us.
all the best,
Jonathan