Does image‐based field work have more to gain from extending or from rejecting scientific realism? An essay in review