Measurement of maintainability early in the software development life cycle, especially during the design phase, may aid designers in incorporating necessary improvements and adjustments to enhance the maintainability of the completed product. In order to demonstrate the importance and necessity of software maintainability during the design phase, this paper expands the (MEMOOD) metrics model, which estimates the maintainability of class diagrams in terms of their understandability and modifiability, to create a multivariate linear model called "Maintainability Estimation Framework and Metrics for Object Oriented Software (MEFOOS)" during the design phase. Thus, class diagrams' maintainability is estimated in terms of their understandability, modifiability, and analyzability in the developed model. This study attempts, as a first effort, to create a relationship between object-oriented design features and maintainability elements analyzability, understandability, and modifiability. In spite of this, the study also created three further multivariate models to measure the class diagram's analyzability, understandability, and modifiability. Finally, the presented models have been empirically validated through the use of relevant statistical tests and reasoning from the context
Research Article
A Metrics-Based Maintainability Estimation Framework for Object Oriented Software in Design Phase
https://doi.org/10.21203/rs.3.rs-2080319/v1
This work is licensed under a CC BY 4.0 License
Version 1
posted
You are reading this latest preprint version
Software maintenance
Software Maintainability
maintainability model
Software Metrics
Software Component
The field of study, application, and research known as software engineering is now crucial. Everyone is putting in a lot of effort to reduce issues and achieve the goal of creating high-quality maintainable software that is delivered on schedule, under budget, and also meets the criteria [1, 2]. Software has become significant to spreading out in almost all region of human endeavor. Making complex programs requires more than just programming ability. There are significant issues with the cost, deadline, post-maintenance, and product quality of software [3, 5]. For the purpose of producing high-quality maintainable software on schedule and within budget, it is essential to put software engineering concepts, strategies, and techniques into practice [5–7]. To do this, we focus in a tightly regulated manner on the software's quality as well as the methodology used to create it [8]. Maintainability has always been a hazy term, making its accurate assessment or evaluation a challenging task. The lack of demonstrated measures for software maintainability is one explanation for this bad management [21]. The ease with which a software system or component can be adjusted to repair errors, increase performance or other features, or adapt to a changed environment is defined as maintainability by IEEE 1990 [4]. The majority of the studies assess maintainability, or more accurately the characteristics that affect it, but only at the source code level. Though measuring maintainability at the source code level is a reasonable way to quantify effort, it causes information to arrive later than planned throughout the software development process [10, 12]. After coding has begun, changing the design to increase maintainability could prove to be exceedingly costly and error prone. Even though early Maintainability estimation in the development life cycle may actually lower the overall development cost. In order to evaluate and, ideally, quantify software maintainability throughout the design phase, this document offers a road map to industry professionals and study [9]. The integration of Maintainability into the development life cycle has been suggested using a prescriptive approach. In terms of software products' maintainability, it can serve as a point of reference.
Thus, our research focus is on one hand, to gauge the viability of system maintainability at the early stages of design, to direct the development of software. Secondly, to build approaches that accurately estimate software maintainability based on design artifacts of class diagrams. In addition to providing some pertinent information about class diagram maintainability, this paper does a thorough review of the maintainability of object-oriented software. The model developed in this paper is an extension of the MEMOOD model and estimates the maintainability of class diagram in terms of their understandability, modifiability, and analyzability. These attributes of the class diagrams are then quantified in terms of object-oriented design metrics calculated from respective class diagram in the model.
Various maintainability prediction models have been put forth in the literature during the past 20 years. The metrics from the coding and design phases are used by some models to forecast maintainability, whereas the metrics at the design level are the only ones they use [13]. We analyze a few of those pieces below.
2.1 Maintainability Measurement Frameworks
[22] presented a maintainability model of class diagram utilizing three strategies viz. Discriminate technique, Weighted-Score-Level procedure, Weighted-Predicted-Level procedure and two sub-attributes of maintainability: understandability and modifiability to assess maintainability model. These attributes are controlled by factors, which may be straightforward or indirectly measured using four software metrics i.e., McCabe’s Cyclomatic Complexity metrics, Reuse Frequency, Regularity, and Halstead's Volume metrics [14]. [15] proposed MI (Maintainability Index) that objectively determines the maintainability of OO software components based on program source code status. [16] suggested a way of scaling OO source-code metrics onto the sub-attributes of maintainability. [17] created a multivariate linear Maintainability Estimation Model (MEMOOD). In terms of understandability and modifiability, the model calculates the maintainability of a UML class diagram. [18], used number of adjustment requests, mean effort per change request and sort of remedy to look at maintainability. In another research [19], presented a maintainability model that groups software modules as ‘not difficult to maintain’ and ‘difficult to maintain’. The model assists developers to select out the modules that are problematic or difficult to take care of, before integrating them. [20], built a maintainability framework by employing polynomial linear regressions. However, this model is implemented distinctly for procedural software and thus not suitable for object-oriented software. [23], developed four frameworks that relate maintainability assessment factors like understandability time and modifiability completeness with structural complexity and metrics of OO component (UML class diagram). However, neither of the four frameworks evaluate the maintainability of OO component itself. [22] provided an Object-Oriented Design Maintainability Estimation Model with Metrics (MOOD). During this study the author discusses the MOOD metrics model for Object-Oriented software, that's designed to assess the maintainability of UML class diagram in terms of testability, understandability, and modifiability. [17] proposed a MEMOOD model, which provide an opportunity to improve the maintainability or understandability of class diagram and consequently the maintainability in final software. [24] describe that the compound MEMOOD model is better the MEMOOD model to determine the maintainability of class diagram in terms of their understandability, modifiability, scalability, and level of complexity. In the work done by [25] offered a novel metric suite to overwhelm the shortages and redefine the relationship amongst design metrics through software maintainability in data intensive applications. [26] evaluated and authenticated the model for software maintainability based on quality factors flexibility and extendibility. The outcomes they arrived stood important but by other factors newer models for maintainability with better-quality outcomes could be proposed. [27] in their work stressed the vast level of cost saving in software by understanding the significance of maintainability, and recommended replies to queries of decision concerning what portions of software to be reused, what portions to be redeveloped, the theoretical valuation of effort requisite to do so and thus giving pointers as how to decrease overall budgets.
In the majority of these studies, maintainability estimation is dependent on the actions done after the coding phase, even though different researchers have offered alternative models for maintainability estimation, according to a review of the literature. As a result, maintainability projections are produced in the later phases of the software development life cycle, making it very challenging to increase maintainability at that point. This work created a multivariate linear model that measures the maintainability of class diagrams in light of the aforementioned important findings.
According to a systematic literature review, there is disagreement among researchers and practitioners about taking maintainability factors into account when evaluating the maintainability of object-oriented software in general and during the design phase [28]. The issue is determining which of those traits should be considered when determining maintainability. [29] has stated at some point in this regard that software entities have many quantifiable properties, and that if all these traits are considered, there are simply too many metrics that will cause more harm than good. As a result, an effective maintainability estimation framework should have as few characteristics as possible while still being capable of evaluating all maintainability elements. This implies that insignificant and intervening features should be excluded from this type of framework [35]. Maintainability of an object-oriented software design is affected by several factors, in which Understandability, Modifiability, and Analyzability is taken as a major factor affecting software maintainability by [12].
Table 1 concludes with a consolidated table for the maintainability factors recognized by area experts. The table clearly shows that the key maintainability factors that we believe influence the maintainability of OO software are Understandability, Modifiability, and Analyzability.
| Maintainability Factors → | Modifiability | Understandability | Stability | Simplicity | Analyzability | Conciseness | Modularity | Testability | Reusability |
|---|---|---|---|---|---|---|---|---|---|
| Study/Source ↓ | |||||||||
| McCall [30] | √ | √ | √ | √ | √ | √ | |||
| Boehm [31] | √ | √ | √ | √ | |||||
| ISO 9126 [32] | √ | √ | √ | √ | √ | √ | |||
| FURPS [33] | √ | √ | √ | ||||||
| Fuzzy Model [34] | √ | √ | √ | √ | |||||
| Rizvi et al. [13] | √ | √ | √ | √ | |||||
| Wiebe et al. [35] | √ | √ | √ |
Object oriented design properties overcome the negative aspect of procedure-oriented design [30, 34]. In order to design the Software through an object-oriented approach, the three essential properties are considerably being used i.e., encapsulation, inheritance, and coupling. Object oriented design properties that have positive impact on maintainability evaluation has been identified and consolidated chart for the same is given in Table 2.
| Design Properties → | Cohesion | Coupling | Encapsulation | Inheritance | Polymorphism |
|---|---|---|---|---|---|
| Source/Study ↓ | |||||
| Gregor et al. (1996) | √ | √ | √ | ||
| Bruce & Shi (1998) | √ | √ | √ | √ | √ |
| B. Pettichord (2002) | √ | √ | √ | ||
| Baudry et al. (2002) | √ | √ | √ | ||
| M Bruntik (2004) | √ | √ | |||
| S. Mouchawrab (2005) | √ | √ | √ | ||
| E Mulo (2007) | √ | √ | √ | √ | |
| Sujata et al. (2011) | √ | √ | |||
| P. Malla et al. (2012) | √ | √ | √ | √ | |
| Nikfard et al. (2013) | √ | √ | √ | √ |
Major maintainability design features mentioned in Table 2, above are often measured by some metrics available in literature [23, 33, 34]. A large portion of those metrics are recognized by specialists on ‘significant utilizations and prominence’ and by scholarly specialists on experimental (post development) validation. But to keep the study simple we have suggested the few basic but popular metrics amongst maintainability researchers that we believe might be applied within the design phase of OO software. The following is an initial recommendation for these metrics.
| Metrics name | Metric Definition |
|---|---|
| Size/Number of the Classes (NC) | The total number of classes |
| Number of attribute (NA) | Total Number of attributes |
| Number of Methods (NM) | Total Number of Methods |
| Number of Association (NAssoc) | The total number of associations |
| Aggregation Number (NAgg) | A class diagram's total number of aggregation relationships. |
| Number of Dependencies (NDep) | The total number of dependencies |
| Number of Generalization (NGen) | The total number of Generalization relationships in a class diagram. |
| Aggregation Number (NAgg) | A class diagram's total number of aggregation relationships. |
| The number of hierarchies of generalization (NAggH) | A class diagram's total number of Generalization Hierarchies. |
| Maximum DIT (MaxDIT) | It represents the highest DIT value obtained for each class in the class diagram. The DIT value for class is the shortest path from the class to the tree's root. |
| Maximum HAGG (MaxHagg) | It represents the highest Hagg value obtained for each class in the class diagram. The longest path from the class to the root of the leaves is the HAGG value for that class. |
Following the successful conclusion of the literature review, there are a number of significant explanations that can be listed.
-
Software quality, customer satisfaction, and overall cost, time, and effort of rework might all be considerably improved if software maintainability were to be assessed at an early stage, i.e., the design phase, in the software development process.
-
A minimal set of maintainability variables for the object-oriented development process must be identified in order to reduce the work required to measure the maintainability of object-oriented design and have a positive impact on maintainability estimation.
-
After recognizing the properties of object-oriented software, a set of maintainability criteria acceptable for the design phase should be determined.
-
Additionally, because metric selection is a crucial stage in measuring the maintainability of objects-oriented architecture, maintainability metrics must be chosen throughout the design phase.
This paper proposes a multivariate linear regression for the establishment of our maintainability estimation framework for OO software components by extending the MEMOOD metrics model with one new factor: analyzability. When discussing the software's maintainability, we are concerned with analyzability, or how simple it is for us to find the components we wish to fix or improve. Hence in our study we believe analyzability is a critical factor that when integrated to existing MEMOOD model gives us a reliable measure for maintainability. In view of this new extra factor, we will build our extended maintainability framework/model, which will help in the long run to reduce the maintenance cost of software project.
7.1 MEMOOD Metrics Model
7.1.1 Introduction
[28] designed MEMOOD metrics which are basically OO design metrics for the enhancement of OO software quality. Figure 1 below shows the existing maintainability model by [28] based on MEMOOD metrics. The maintainability model's quantification procedure is shown in Fig. 1 below (MEMOOD). Class diagram understandability and modifiability are measured using two ('NC' and 'NGenH') and five ('NC', 'NGen', 'NGenH', 'NAggH', and 'MaxDIT') class diagram metrics, respectively. While maintainability is assessed using understandability and modifiability.
7.1.2 Proposed extended model based on MEMOOD model
A new framework for measuring the maintainability of OO software is proposed in this paper by extending the MEMOOD metrics model. We introduce one new factors analyzability. These factors enhance maintainability of object-oriented design components up to higher level. In this extended framework we correlate the prevailing and new factors i.e., “Modifiability, Understandability, and Analyzability” as the major maintainability attributes with different OO Design factors that influence them. The following are the components of our suggested framework: (i) Key attributes that influence maintainability. (ii) measurable OO structures that decide each maintainability factor. (iii) measurement metrics and (iv) the ultimate maintainability calculation model.
The study used data from two different sources. Kiewkanya et al [22]'s dataset was used in one dataset. It includes data acquired from a controlled experiment on understandability, modifiability, and maintainability. This dataset was utilized in regression analysis to create the maintainability model (MEMOOD), with maintainability acting as the dependent variable and understandability and modifiability as independent variables. Genero et al [21]'s dataset was used in another dataset. It includes understanding and modifiability scores as well as eleven metrics (given in Table 3) derived from 28 class diagrams. This dataset was utilized in the study to build four different multivariate linear regression models for the class diagram's understandability, modifiability, and analyzability, using the metrics of the class diagram as independent variables.
The prerequisite for a precise maintainability measurement model is the measurement of class diagram analyzability, understandability, and modifiability. The study has therefore created three models for Analyzability, Understandability, and Modifiability before establishing MEFOOS. Multivariate linear regression model (1) has been chosen to be the setup model for all three models.
| Y = β0 + β1* X1 + β2 * X2 +………… + βn * Xn + ε | (1) |
Where
Y is the projected value of the dependent variable while X1, X2 …, Xn are independent variables. β1, β2, …,βn is the regression coefficient of independent variables, ε model error (i.e., the degree of variation in our estimate of ) and β0 is the y-intercept (Y value when every other parameter is set to 0).
9.1 Understandability Model
Metrics listed in Table 3 will serve as independent variables while Understandability will be used as the dependent variable to create a multivariate model for class diagram Understandability. The necessary information utilized to create the Understandability measurement model is sourced from [3]. The association in the middle of maintainability factors and object-oriented characteristics has been established as illustrated in Eq. 2. As per the mapping, Metrics ‘NC’, ‘NGenH’ are selected from Table 3 as independent variable to set up the Understandability measurement model. The coefficient values are obtained using SPSS, and the Understandability model is formulated as follows:
| Understandability = 1.617 − 0.016 * NC + 0.377 * NGenH | (2) |
Where NGenH stands for "Number of Generalization Hierarchies" and NC stands for "Number of Classes." According to (2), "NC" is directly proportional to "Understandability of Class Diagram," but "NGenH" is inversely proportional to "Understandability of Class Diagram."
9.1.1 Statistical Significance of the Model and Independent Variables
It is clear from the numbers in Tables 3, 4, and 5 that the Understandability model is statistically significant at a level of confidence more than 99%. Additionally, the values of R2 and Adjusted R2 are satisfactory. The metrics "NC" and "NGenH" involved in (2) are both statistically significant.
| Model | Unstandardized Coefficients | Standardized Coefficients | ||
|---|---|---|---|---|
| B | Std. Error | Beta | ||
| 1 | (Constant) | 1.617 | 0.367 | |
| NC | -0.016 | 0.019 | -0.304 | |
| NGenH | 0.377 | 0.222 | 1.066 | |
| a. Dependent Variable: Understandability | ||||
| Model | R | R2 | Adjusted R2 | Std. Error of the estimate |
|---|---|---|---|---|
| 1 | 0.987a | 0.975 | 0.952 | 0.38277 |
| a. Predictors: (Constants), NC, NGenH. | ||||
| Model | Sum of Squares | df | Mean Square | F | Significance. | |
|---|---|---|---|---|---|---|
| 1 | Regression | 55.045 | 2 | 27.526 | 120.879 | 0.006a |
| Residual | 4.098 | 18 | 0.228 | |||
| Total | 55.143 | 20 | ||||
| a. Predictors: (Constants), NC, NGenH | ||||||
| b. Dependent Variable: Understandability | ||||||
9.2 Modifiability Model
The next step is to create a similar model for Modifiability after building a model for Understandability. Data from [22] was utilized to create the modifiability model. Eq. 3 illustrates the proven relationship between object-oriented qualities and maintainability factors. Metrics "NC", "NGenH", "NGen", "NAggH" and "MaxDIT" are chosen from Table 3 as independent variables to construct the Modifiability measurement model in accordance with the mapping. The coefficient values are obtained using SPSS, and the modifiability model is formulated as follows:
| Modifiability = 1.088 + 0.283 * NC – 0.079 * NGen – 0.616* NGenH – 0.696 * NAggH + 0.170 * MaxDIT | (3) |
The "Number of Classes" (NC), "Number of Generalizations" (NGen), "Number of Aggregation Hierarchies" (NAggH), "Number of Generalization Hierarchies" (NGenH), and "Maximum DIT" (MaxDIT) are used in this equation. According to the model, "Number of Classes" and "Maximum DIT" are directly correlated with "Modifiability of Class Diagram," whereas "NGen" and "Number of Generalization and Aggregation Hierarchies" are inversely correlated.
9.2.1 Statistical Significance of the Model
Observing the significance (p-value) for the F-test in the last column of Analysis of Variance (Table 6), it can be concluded that the modifiability model (3) is statistically significant at a confidence level of more than 95%.
| Model | Sum of Squares | df | Mean Square | F | Significance. | |
|---|---|---|---|---|---|---|
| 1 | Regression | 60.237 | 5 | 12.047 | 114.886 | 0.079a |
| Residual | 1.573 | 15 | 0.105 | |||
| Total | 61.810 | 20 | ||||
| a. Predictors: (Constants), MaxDIT, NGen, NGenH, NC, NAggH. | ||||||
| b. Dependent Variable: Modifiability | ||||||
Also, the value of R2 (Coefficient of Determination) and Adjusted R2 in the Table 7, is also very encouraging. As, it refers to the percentage or proportion of the total variance in modifiability by all the five metrics (independent variables) participating in the model (3).
| Model | R | R2 | Adjusted R2 | Std. Error of the estimate |
|---|---|---|---|---|
| 1 | 0. 980a | 0.960 | 0.879 | 0.324 |
| a. Predictors: (Constants), MaxDIT, NGen, NGenH, NC, NAggH. | ||||
9.2.2 Statistical Significance of Independent Variables
It can be noticed from the end column of Table 8, that all of the five metrics participating in the model is statistically significant at a significance level of 0.05 (equivalent to a confidence level of 95%).
| Model | Unstandardized Coefficients | Standardized Coefficients | ||
|---|---|---|---|---|
| B | Std. Error | Beta | ||
| 1 | (Constant) | 1.088 | 0.224 | |
| NC | 0.283 | 0.077 | 1.341 | |
| NAssoc | -0.079 | 0.048 | -0.204 | |
| NGenH | -0.616 | 0.241 | -0.316 | |
| NAggH | -0.696 | 0.192 | -0.255 | |
| MaxDIT | 0.170 | 0.104 | 0.122 | |
| a. Dependent Variable: Modifiability | ||||
9.3 Analyzability Model
The next step is to create a similar model for Analyzability after building a model for Understandability and Modifiability. Data from [22] was utilized to create the modifiability model. Eq. 4 illustrates the proven relationship between object-oriented qualities and maintainability factors. Metrics "NC", "NGenH" and "MaxDIT" are chosen from Table 3 as independent variables to construct the Analyzability measurement model in accordance with the mapping. The coefficient values are obtained using SPSS, and the analyzability model is formulated as follows:
| Analyzability = 1.617–0.016 * NC + 0.377 * NGenH + 0.671 * MaxDIT | (4) |
Where, NC is the ‘Number of Classes’, NGenH is ‘Number of Generalization Hierarchies’ and MaxDIT is Maximum DIT. From the model it can be interpreted that modifiability of class diagram is directly proportional to ‘Number of Classes’ and ‘Maximum DIT’, while ‘NGenH’ is inversely proportional to analyzability of class diagram.
9.3.1 Statistical Significance of the Model and Independent Variables
According to the results of Tables 9, 10, and 11, the Analyzability model is statistically significant at a level of confidence more than 99%, and the values of R2 and Adjusted R2 are satisfactory as well. Both of the three metrics involved in (4)—NC, NGenH, and MaxDIT—are statistically significant.
| Model | Unstandardized Coefficients | Standardized Coefficients | ||
|---|---|---|---|---|
| B | Std. Error | Beta | ||
| 1 | (Constant) | 1.617 | 0.367 | |
| NC | -0.016 | 0.018 | -0.304 | |
| NGenH | 0.377 | 0.50 | 0.546 | |
| MaxDIT | 0.671 | 0.222 | 1.065 | |
| a. Dependent Variable: Analyzability | ||||
| Model | R | R2 | Adjusted R2 | Std. Error of the estimate |
|---|---|---|---|---|
| 1 | 0.987a | 0.975 | 0.952 | 0.38277 |
| a. Predictors: (Constants), NC, NGenH, MaxDIT | ||||
| Model | Sum of Squares | df | Mean Square | F | Significance. | |
|---|---|---|---|---|---|---|
| 1 | Regression | 55.045 | 2 | 27.526 | 120.879 | 0.006a |
| Residual | 4.098 | 18 | 0.228 | |||
| Total | 55.143 | 20 | ||||
| a. Predictors: (Constants), NC, NGenH, MaxDIT | ||||||
| b. Dependent Variable: Analyzability | ||||||
9.4 Maintainability Estimation Model
It is crucial to ensure the proper relationship between maintainability, understandability, modifiability, and analyzability of class diagrams before constructing the maintainability model. Table 12 displays the correlation coefficients between them. The correlation values show a substantial association between maintainability and both Understandability, Modifiability and Analyzability.
| Maintainability | Understandability | Modifiability | Analyzability | ||||
|---|---|---|---|---|---|---|---|
| Pearson Correlation | Maintainability | 1.000 | 0.792 | 0.822 | 0.400 | ||
| Understandability | 0.792 | 1.000 | 0.792 | 0.792 | |||
| Modifiability | 0.822 | 0.757 | 1.000 | 0.822 | |||
| Analyzability | 0.400 | 0.567 | 0.400 | 1.000 | |||
| Maintainability = 2.168–0.313 * Understandability – 0.938 * Modifiability + 1.043 * Analyzability. | (5) | ||||||
| Model | Unstandardized Coefficients | Standardized Coefficients | ||
|---|---|---|---|---|
| B | Std. Error | Beta | ||
| 1 | (Constant) | 2.168 | 0.599 | |
| Understandability | -0.313 | 1.156 | -1.220 | |
| Modifiability | -0.938 | 0.745 | -4.172 | |
| Analyzability | 1.043 | 1.543 | 4.896 | |
| a. Dependent Variable: Maintainability | ||||
| Model | R | R2 | Adjusted R2 | Std. Error of the estimate |
|---|---|---|---|---|
| 1 | 0.879a | 0.773 | 0.091 | 0.48908 |
| a. Predictors: (Constants), Understandability, Modifiability, Analyzability | ||||
| Model | Sum of Squares | df | Mean Square | F | Significance. | |
|---|---|---|---|---|---|---|
| 1 | Regression | 0.813 | 3 | 0.271 | 1.133 | 0.583a |
| Residual | 0.239 | 1 | 0.239 | |||
| Total | 1.239 | 4 | ||||
| a. Predictors: (Constants), Understandability, Modifiability, Analyzability | ||||||
| b. Dependent Variable: Maintainability | ||||||
This section of the paper assesses, how well the models developed in the previous section are able to quantify the modifiability, understandability, analyzability, and maintainability of new class diagrams those have not participated in the development of the models. 75% of the available data is being used for developing the models, while remaining 25% used for validating them. As long as validation process is concerned, the modifiability, understandability, analyzability, and maintainability have been calculated using respective model’s equation. Subsequently Pearson’s correlation coefficient is being calculated between the calculated (using model equation) and actual (already known) values of modifiability, understandability, analyzability, and maintainability.
|
Projects |
Maintainability Ranking |
d |
∑d2 |
rs |
rs > 0. 7324 |
|
|---|---|---|---|---|---|---|
|
Calculated Ranking |
Known Ranking |
|||||
|
P1 |
15 |
1 |
14 |
196 |
0.711 |
✓ |
|
P2 |
14 |
2 |
12 |
144 |
0.788 |
✓ |
|
P3 |
3 |
3 |
0 |
0 |
1 |
✓ |
|
P4 |
12 |
4 |
8 |
64 |
0.905 |
✓ |
|
P5 |
5 |
5 |
0 |
0 |
1 |
✓ |
|
P6 |
6 |
6 |
0 |
0 |
1 |
✓ |
|
P7 |
7 |
9 |
-2 |
4 |
0.994 |
✓ |
|
P8 |
1 |
11 |
-10 |
100 |
0.852 |
✓ |
|
P9 |
13 |
7 |
6 |
36 |
0.947 |
✓ |
|
P10 |
9 |
12 |
-3 |
9 |
0.986 |
✓ |
|
P11 |
8 |
13 |
5 |
25 |
0.963 |
✓ |
|
P12 |
10 |
14 |
4 |
16 |
0.976 |
✓ |
|
P13 |
4 |
15 |
-11 |
121 |
0.822 |
✓ |
|
P14 |
11 |
8 |
3 |
9 |
0.986 |
✓ |
| rs > 0.7324 means significant results. | ||||||
Known values and their calculated values were correlated using a constructed model, and Spearman's Coefficient of Correlation (rs) was utilized to validate this relationship. Finding the degree of correlation between two variables is done through the process of rank correlation. Thus, it is possible to conclude without losing any generality that the maintainability estimation model measurements are in fact reliable, significant, and useful. According to the following formula below, the "rs" was determined.
Where
d = difference between Maintainability's "Known Ranking" and "Calculated Ranking".
n = numbers of pairs of rankings used in the experiment.
The study has developed four models to compute Understandability, Analyzability, Modifiability, and Maintainability of the class diagrams. Maintainability model measures the maintainability of object-oriented design in terms of their Understandability, Analyzability and Modifiability. All the models have been developed using the process of multiple linear regressions and validates the measuring ability of maintainability model. The practical validation on the maintainability measurement model shows that developed models are highly consistent, up to standard and significant. The values of Understandability, Analyzability, Modifiability and Maintainability are of instant use in the software development life cycle. In order to manage or at least reduce future maintenance costs, these values may aid software designers in reviewing their designs and taking the needed remedial actions early in the development life cycle.
This information may also be used by the maintenance team to decide which module to concentrate on during maintenance. By completing a bigger scale study with a range of industrial projects across many disciplines, a more advanced maintainability estimation model may be created in the future. In the future, a more generic maintainability estimating model may be built, however the maintainability estimation model (MEFOOS) created in this study concentrates on the object-oriented paradigm. Future studies may also focus on measuring additional quality attributes suggested by ISO 9126, such as reusability, portability, reliability, testability, etc. A maintainability index can be created in addition to these, which could aid the software industry in project rating.
Authors Contribution
All the three authors have contributed equally to the work. All authors read and approved the final manuscript.
Data Availability
Research involving Human Participants and/or Animals: This article does not contain any studies with human participants or animals performed by any of the authors.
Informed consent: Informed consent was obtained from all individual participants included in the study.
Funding Declaration
No funding was pursued or granted for this paper.
Conflict of Interest
The authors declare that they have no potential conflict of interest in relation to the study in this paper
- M. Dagpinar and J. Jahnke, “Predicting Maintainability with Object- Oriented Metrics – an Empirical Comparison,” Proc. 10th Working Conference on Reverse Engineering (WCRE‟03), 13–17 Nov. 2003, pp. 155–164, 2003.
- Abdullah, Dr, Reena Srivastava, and M. H. Khan. “Modifiability: A Key Factor to Testability”, International Journal of Advanced Information Science and Technology, Vol. 26, No.26, Pages 62–71 June 2014.
- M. Genero, E. Manso, A. Visaggio, and M. Piattini, “Building Measure-Based Prediction Models for UML Class Diagram Maintainability,” Journal of Empirical Software Engineering, vol.12, no. 5, pp. 517–549, 2007.
- International Organization for Standardization. ISO/IEC 9126-1: Software engineering - Product quality - Part 1: Quality model, 2001.
- J.H. Hayes and L Zhao, “Maintainability Prediction: a Regression Analysis of Measures of Evolving Systems,” Proc. 21st IEEE International Conference on Software Maintenance, 26–29 Sept. 2005, pp. 601–604, 2005.
- Abdullah, Dr, M. H. Khan, and Reena Srivastava. “Testability Measurement Model for Object Oriented Design (TMMOOD).” International Journal of Computer Science & Information Technology (IJCSIT) Vol. 7, No 1, February 2015, DOI:10.5121/ijcsit.2015.7115.
- M.O. Elish and K.O. Elish, “Application of TreeNet in Predicting Object-Oriented Software Maintainability: a Comparative Study,”Proc. of European Conference on Software Maintenance and Reengineering (CSMR‟09), 24–27 Mar., 2009, pp. 69–78,2009.
- IEEE Press, “IEEE Standard Glossary of Software Engineering Technology,” ANSI/IEEE Standard 610.12–1990, 1990.
- M. Polo, M. Piattini, and F. Ruiz, “Using Code Metrics to Predict Maintenance of Legacy Programs: A Case Study,” Proc. of International Conference on Software Maintenance, ICSM 2001, pp. 202–208, IEEE Computer Society, Florence Italy, 2001.
- Abdullah, Dr, Reena Srivastava, and M. H. Khan. "Testability Measurement Framework: Design Phase Perspective."International Journal of Advanced Research in Computer and Communication Engineering Vol. 3, Issue 11, Pages 8573–8576 November 2014.
- Jagdish Bansiya, “A Hierarchical Model for Object Oriented Design Quality Assessment”, IEEE Transaction of Software Engineering, Volume 28, No. 1, January 2002, and pp: 4–17
- J.H. Hayes, S.C. Patel, and L. Zhao, “A Metrics-Based Software Maintenance Effort Model,” Proc. 8th European Conference on Software Maintenance and Reengineering (CSMR'04), 24–26 Mar. 2004, pp. 254–258, IEEE Computer Society, 2004.
- S.W.A. Rizvi and R.A. Khan, “A Critical Review on Software Maintainability Models,” Proc. of the National Conference on Cutting Edge Computer and Electronics Technologies, 14–16 Feb. 2009, pp. 144–148, Pantnagar, India, 2009.
- Budhija, N., Singh, B., & Ahuja, P. S. (2013, January). International Journal of Advanced Research
- P.W. Oman and J.R. Hagemeister, “Construction and Testing of Polynomials Predicting Software Maintainability,” Journal of Systems and Software, vol. 24, no. 3, pp. 251–266, 1994.
- Caldiera, G., &Basili, V. R. (1991). Identifying and Qualifying Software Components. 24.
- Rizvi, S. W., & Khan, R. A. (2010, APRIL). Maintainability Estimation Model for Object-Oriented Software in Design Phase (MEMOOD). JOURNAL OF COMPUTING, 2(4), 26–32.
- M. Polo, M. Piattini, and F. Ruiz, “Using Code Metrics to Predict Maintenance of Legacy Programs: a Case Study,” Proc. of International Conference on Software Maintenance, ICSM 2001, pp. 202–208, IEEE Computer Society, Florence Italy, 2001.
- J.H. Hayes and L Zhao, “Maintainability Prediction: aRegression Analysis of Measures of Evolving Systems,” Proc. 21st IEEE International Conference on Software Maintenance, 26–29 Sept. 2005, pp. 601–604, 2005.
- S. Muthanna, K. Kontogiannis, K. Ponnambalam, and B. Stacey, “A Maintainability Model for Industrial Software Systems Using Design Level Metrics,” Proc. 7th Working Conference on Reverse Engineering (WCRE’00), 23–25 Nov., 2000, pp. 248–256, Brisbane, Australia, 2000.
- M. Genero, E. Manso, and G. Cantone, “Building UML Class Diagram Maintainability Prediction Models Based on Early Metrics,” Proc. 9th International Symposium on Software Metrics (METRICS'03), 3–5Sept. 2003, pp. 263–275, 2003.
- M. Kiewkanya, N. Jindasawat, and P. Muenchaisri, “A Methodology for Constructing Maintainability Model of Object-Oriented Design,Proc. 4th International Conference on Quality Software, 8–9 Sept., 2004, pp. 206–213. IEEE Computer Society, 2004.
- Kiranjit Kaur Sami Anand, A Maintainability Estimation Model and Metrics for Object-Oriented Design (MOOD), ISSN: 2278–1323 International Journal of Advanced Research in Computer Engineering & Technology (IJARCET) Volume 2, No 5, May 2013
- Gautam C., kang S.S., “Comparison and Implementation of Compound MEMOOD MODEL and MEMOOD MODEL”, International journal of computer science and information technologies, pp 2394–2398, 2011.
- R. Malhotra and A. Chug, Application of evolutionary algorithms for software maintainability prediction using object-oriented metric, in 8th Int. Conf. Bio-Inspired In- formation and Communications Technologies, 2014.
- Rajendra Kumar and Dhanda N, Maintainability Measurement Model for Object-Oriented Design, International Journal of Advanced Research in Computer and Communication Engineering, Vol. 4, Issue 5, May 2015.
- Celia Chen, and Lin Shi, Why Is It Important to Measure Maintainability and What Are the Best Ways to Do It?, IEEE/ACM 39th
- Hristov, D., Hummel, O., Huq, M., &Janjic, W. (2012). Structuring Software Reusability Metrics. International Conference on Software Engineering Advances. IARIA.
- Westfall, L. (2005). 12 Steps to Useful Software Metrics, the Westfall team: Retrieved from: http://www.westfallteam.com/Papers/12_steps_paper.pdf
- McCall, J.A., Richards, P.K., and Walters, G.F., (1977) “Factors in Software Quality”, RADC TR-77-369, Vols I, II, III, US Rome Air Development Center Reports.
- Boehm, B. W., Brown, J. R., Kaspar, H., Lipow, M., McLeod, G., and Merritt, M., (1978) Characteristics of Software Quality, North Holland.
- ISO 9126-1 Software Engineering - Product Quality - Part 1: Quality Model, 2001.
- Grady, Robert, Caswell, Deborah (1987), Software Metrics: Establishing a Company- wide Program. Prentice Hall. pp. p. 159. ISBN 0138218447.
- K.K. Aggarwal, Y. Singh, P. Chandra and M. Puri, “Measurement of Software Maintainability Using a Fuzzy Model”, Journal of Computer Sciences, vol. 1, no.4, pp. 538–542, 2005 ISSN 1549–3636 © 2005 Science Publications.
- Hordijk, Wiebe, and Roel Wieringa. "Surveying the factors that influence maintainability: research design." ACM SIGSOFT Software Engineering Notes. Vol. 30. No. 5. ACM, 2005.
No competing interests reported.