I work with fibre optic components and one of the specifications that I typically find for telecordia class components is a power handling capability of say 300-500mW. My first reaction is that a power that low must be an average power but then if you have a high power peak then why wouldn't that burn a component. In my mind I consider the absorption and heat transfer of the material and deduce that if a pulse is sufficiently narrow in time then a high peak can perhaps fail to be absorbed or rapidly dispersed through the medium to such an extent to which the impact my not be severe. Then I switch to the idea that perhaps the average power is the quantity to focus on because it's referring to, on average, how much power is incident at all time even if the heat is being transferred away by the natural properties of the material. This argument goes round in circles and I'm not fully convinced which power definition is the best to keep in mind.