Muddling through in innovation — On incremental failure in developing an engine