Scientific software must be personalized for dissimilar execution environments, problem sets, and existing resources to make sure its competence and consistency. Even though adaptation patterns be able to be found in a extensive percentage of new scientific applications, the conventional scientific software stack be deficient in the sufficient adaptation concept and tools. As a consequence, scientific programmers physically implement ad-hoc solutions that are tough to retain and reuse. In this paper, illustrate the experimental evaluation of the level of inheritance in five object-oriented systems. The systems considered differ in together its size and application domain. A consequence from our analysis mutually with added new investigations appears to support the thesis that inheritance is used moreover sparingly or incorrectly. Statistical relationship among four inheritance metrics and a position of reliant variables (non-comment source lines, software understandability, known errors and error density) present evidence for this claim. It is also not clear that systems with the use of inheritance will essentially be more maintainable than those that do not. The data examined from two of our systems recommends that deeper inheritance trees are characteristics of systems which are harder to understand and (by implication) harder to preserve. By examining, why? This may be the case, and suggest ways of remedying this state of affairs.