merge m things with n (minimum average)

[Michel Schellekens <m.schellekens@cs.ucc.ie>]

I guess a second question is the following: when one looks at the paper
``New algorithms for selection'', 
it does give an optimal algorithm for selecting the k largest values
from n, but the authors also say that their 
algorithm only works for values of k which are small compared to n. It
seems this is not sufficient to 
deduce information on the general average time required to merge? If it
is, I'd appreciate a hint or 
a pointer to another result in the paper which might give the conclusive
answer.

Thanks,
Michel Schellekens.

-- 
Dr. M. P. Schellekens                | Tel. + 353 (0)21 90 3083
National University of Ireland, Cork | Fax. + 353 (0)21 90 3113
Department of Computer Science       | Email: m.schellekens@cs.ucc.ie
Cork, Ireland                        | WWW: http://www.cs.ucc.ie/~mpcs/