[DTownSMR]

For many years I have used (and implemented in numerous tools) the convention: transpose up produces sharps, down produces flats.

So a transposition of +2 of A - E7sus - F#m - D will produce B - F#7sus - G#m - E, while a transposition -10 will produce B - Gb7sus - Abm - E.

This is consistent, easy to implement, and easy to understand/explain.

Programatically I get what you're saying but I'm not sure I get WHY? i.e. what is the use case for supporting Gb7sus & Abm in the key of B?!? The "issue" as I see it is that I've selected a new KEY and simply expect the correct chord names to be used. I don't want to have to think, or even know, which is "right". (my two years of theory are long past and I've forgotten most of it anyway! lol)  In other words, If we're all used to seeing chords in the key of X with certain accidentals then why not have the program know that...that's the usefulness of software. :-)

Perhaps the notion of "up for sharp & down for flat" makes sense in picking between the key of F# & Gb but even then I'm not sure that would make the best UI experience. I would have to mock some prototypes up say with any confidence.