I think that the suggestion of using additionalPayment does not work, because the additionalPayments are deterministic – the cashflow will occur with specified magnitude on the specified date – whereas the option payoff is contingent on attaining the strike – this fact is unknown until the fixing. Second, a cap/floor digital payoff is usually expressed as a rate, applied to the notional for the applicable coverage (duration) and day-count convention, and this is not supported by a “payment” representation. Philip Leach at DTCC recently produced a propoal on digital cap/floors (see FpML Issue #900) and I have been working on a response.