??so what about how general theme of "non-degenerate inner product produces canonical flag generic to given flag" (and so forth.... ???relationship of flag variety to conjugacy class and/or coadjoint orbit, and so forth ... ???) relates to ... ???something about "algebraic geometry issues" ?? ... or something... ??? ???something about whether some version of this idea works in an "ag" context ??... ??something about that bit about how... ??"gram-schmidt fails in complex case ..." ... ??or something??? ... i'm probably misremembering how that went; i should have sufficient notes written down somewhere to check up on it ...