Sure, but if we want to hammer out what playflows and affordances we want from a second combo, we are going to have to look at what we like or don't like from said combos.
Starting from the ones already in-game before moving on to more theoretical grounds seems a good plan.
We can, also, look at both other forms of GCD variety and what we want from paced/secondary combos. We needn't "rather than" either out of the picture.
___________
For my part, fixed CD combos seem the least interesting, and also the most frustrating if one has no other way to correct for the drift they can cause.
Similarly, I generally prefer DoT durations to buff durations, because they at least add two further considerations, cleave and the target's remaining lifespan (in seconds' time).
Resource-generation could potentially take the cake, though, if able to introduce some degree of polyrhythm (a shorter use here, a longer use there) instead of being obliged to be used at, in effect, an equally locked interval. (If not, though, it'd typically be the worst of all worlds -- increasing downtime punishment in a way that'd also badly further drift, causing one to entering combat without access to their cooler-looking skills, and generally feeling incredibly constraining.)
Ideally, resource-generated combos should also have some banking available to them, along with some capacity to take up or give slack to counteract would-be drift. Variable length resource-spending combos could probably do this best, so long as there is a good reason to use the combo even when unable to complete it. For instance, if a combo granted some duration of a buff for even just starting it, you'd typically (but ideally, not always) prioritize keeping that buff up over maximizing the ppgcd of that combo by using it only when you can support its full length.
