I've done quite a bit of performance testing, and found that in a majority of cases evaluate() is faster (by far) than doing the same thing in Lotuscript. Lotuscript is always interpreted where Evaluate() calls a formula language processing api which is compiled. In one test case, I compared @password() for creating hash to an actual lotuscript written hash algorithm, and the rest was hundreds of times faster in evaluate.
Another advantage, as in this case, is that you increase the chances of a bug (yours or lotus's) that would result in your two code paths having a different result when you need to perform the same routine in both languages. Less code == better code.
I've done quite a bit of performance testing, and found that in a majority of cases evaluate() is faster (by far) than doing the same thing in Lotuscript. Lotuscript is always interpreted where Evaluate() calls a formula language processing api which is compiled. In one test case, I compared @password() for creating hash to an actual lotuscript written hash algorithm, and the rest was hundreds of times faster in evaluate.
Another advantage, as in this case, is that you increase the chances of a bug (yours or lotus's) that would result in your two code paths having a different result when you need to perform the same routine in both languages. Less code == better code.