Lewis 



Lewis log-type spectrum is equal to the mean squared wave slope and not the 

 mean slope which presumably must be zero just like the mean wave amplitude 

 is considered to be zero. 



There is next, a definite and precise meaning attached to the complex fre- 

 quency response and I suggest that arbitrary interpretations had better be 

 avoided. What the meaning of the now accepted word- response amplitude oper- 

 ator should be, is simply the square of the response amplitude measured in its 

 own units due to a unit amplitude of the excitation, be it wave or wave-induced 

 load. In advocating his non-dimensional procedure. Professor Lewis is forced, 

 on account of the large number of ship responses, to examine and adopt differ- 

 ent parameters which will non-dimensionalize each individual response. Hence, 

 the cause of such confusing statements like "wave slope is more important for 

 pitch motion than wave amplitude." What Professor Lewis means is that if one 

 wants to non-dimensionalize an angular displacement he had better divide by a 

 (dimensionless) angle such as maximum wave slope. Clearly then, because we 

 have many and different "responses" in the ship-system case, non-dimensional- 

 ization is of no real use and only adds to undue complication. 



I also fail to see the legitimacy of multiplying two arbitrarily derived func- 

 tions in order to get a response spectrum, unless these functions indeed repre- 

 sent quantities which specifically relate themselves to the fundamental notions 

 behind the theory of linear systems. The advantage that the author claims is 

 that the effects of ship size can be readily shown. But by size. Professor Lewis 

 limits himself to length only. What about variations in say breadth or draft or 

 water-plane coefficient when the "useful shift" of the curves doesn't take place? 

 Do we have to start all over again with new non-dimensionalizing ? 



An important final point is that in the end of our analysis, we should not be 

 satisfied with simple families of curves. The trends, once established, are 

 only a palliative; the really useful information to the designer is rather numbers 

 like the ones Dr. Ochi has discussed in his paper. The author presumably makes 

 a plea at the end of his paper that we should avoid masses of dusty information. 

 Personally, I can think of no better way to fill drawers than by attempting to col- 

 lect curves for all possible variables. 



The last section of the paper is the most interesting and it is a pity that the 

 author did not amplify the basic problem. As I see it, there are now three 

 things to be done before we can really say that we are incorporating our knowl- 

 edge in ship design. 



The first thing is related to the oceanographers and here we must wait for 

 their answer to the basic question. In a given year (or even better in a period 

 of years) and over a specified ship route what are the sea spectra encountered 

 by a ship and what is their individual time occurrence ? 



The second thing is to determine in numerical terms exactly what ship op- 

 erators mean by unacceptable wetness, untolerable number of slams or unbear- 

 able acceleration. 



216 



