In order to achieve consistent management of design and analysis of engineering objects, data in engineering applications are required to be manipulated in an integrated manner. Engineering data comprise attributes, structural and pictorial information. ADAM uses the relational model as a basis for representing attributes information and structures of engineering data. As for pictorial information, ADAM data definition language provides figure specification parts where pictorial notations of instances are represented in terms of functions in computer graphics. In this paper, we observe engineering data and give some considerations on them. Four essential features of engineering data are discussed. Data definition and manipulation facilities of ADAM are presented. Instances with arguments are introduced for managing complex structures of engineering data and repeated appearances of similar objects. A graphic algorithm is discussed, which allows users to retrieve pictures in any intended details.

An algorithm is given which selects the Kth element in $X + Y$ in $O(n\log n)$ time and $O(n)$ space, where $X + Y$ is the multiset $\{ x_i + y_j | x_i \in X\text{ and } y_j \in Y\} $ for $X = (x_1 ,x_2 , \cdots ,x_n )$ and $Y = (y_1 ,y_2 , \cdots ,y_n )$, n-tuples of real numbers. The results are extended to $\sum_{i = 1}^m {X_i } $ for $m > 2$. There is strong evidence that this more general problem is difficult if m and K may be selected arbitrarily. However, algorithms can be shown which are fast for small K and arbitrary m.