A general economics model of software reuse

Abstract

Ageneralmodelofsoflwai-e reuse economics ispresented. The model provides a framework for making estimates and decisions about the economic desirability of reusing sojiware. It covers costs and productivity, return on investment, and incremental fimding schemes,

Reusing Code from StackOverflow: The Effect on Technical Debt

“…In principle, reuse should enhance the quality of software, primarily because of the increased opportunity it provides for error discovery. Quoting Gaffney and Cruickshank [8], "each time reusable code is used in a new application, an additional opportunity is provided for error discovery and removal". Early empirical studies have provided evidence that systematic reuse can have a positive effect on software quality, especially leading to lower fault-density [9]- [11].…”

Validating pragmatic reuse tasks by leveraging existing test suites

“… Re‐implementation is expensive and does not take advantage of the maturity of tested, debugged, and deployed code . Preplanned reuse is the traditional route , and is found in approaches as varied as object‐oriented inheritance, software components, and software product lines. Although practical in some situations, all these approaches have drawbacks more generally: (i) it is difficult to predict which parts of a system should be developed with future reuse in mind ; (ii) big upfront costs are incurred that are not necessarily recouped by savings over the long‐term; thus, it is economically infeasible to develop all software with future reuse in mind ; and (c) software that has been developed to be reusable still embeds a set of assumptions that will not be valid in all contexts, thereby limiting its ability to be reused as‐is . Refactoring the original system to create better reusable software entities within it is expensive, risky, and not always possible : (i) the developer may not own the original system (even when its license allows them to reuse the source code) so their refactoring efforts could only affect their local copy; (ii) the original system may already be deployed, and so, cannot be readily modified in substantial ways; (iii) the developer of the original system may not be willing to change its design, which could introduce defects into a working system; (iv) there are security implications to sharing code that can be changed as someone else sees fit; and (v) it does not benefit the financial situation of an organization in the short‐term.…”

“…Preplanned reuse is the traditional route , and is found in approaches as varied as object‐oriented inheritance, software components, and software product lines. Although practical in some situations, all these approaches have drawbacks more generally: (i) it is difficult to predict which parts of a system should be developed with future reuse in mind ; (ii) big upfront costs are incurred that are not necessarily recouped by savings over the long‐term; thus, it is economically infeasible to develop all software with future reuse in mind ; and (c) software that has been developed to be reusable still embeds a set of assumptions that will not be valid in all contexts, thereby limiting its ability to be reused as‐is .…”

“…Software reuse has been promoted since the 1960s as a means to improve software quality and to increase development productivity . Classic approaches to software reuse involve predicting which functionality will need to be reused in the future, and investing heavily in packaging that functionality in a reusable form; this can be difficult and expensive , and can still fail to yield the necessary reusable components . Although preplanned reuse has a successful place in industrial practice, many disciplined software reuse tasks involve an ad hoc, pragmatic process of copying segments of source code and modifying these to integrate with a target system ; pragmatic reuse is industrially necessary and beneficial for a variety of organizational and technical reasons .…”

Aligning the economic modeling of software reuse with reuse practices

“…Many reuse economic models have been proposed in the literature, see e.g. [5][6][7][8]18,22,23,28,[30][31][32]34]. The benefits of software reuse are generally defined in terms of the cost difference between the development of applications without reuse and the development of applications with reuse.…”