CN110245869A - Customizable open source software selection method towards different field - Google Patents

Abstract

A customizable open source software selection method for different fields relates to the field of software engineering technology. In order to solve the problem of a lack of a customizable open source software selection method for different fields in the prior art, the method includes, step 1: according to the application field of the software. Customize the metrics, and then collect the metrics data; Step 2: Select a supervisory group; Step 3: Publish the collected metrics data information to the crowdsourcing workers; Step 4: The crowdsourcing workers decide the quality model First The weights of the first layer and the second layer are scored, and the supervision team screens the obtained results and determines whether they meet the screening criteria. If they meet the requirements, go to step five; if they do not meet the requirements, go to step three; step five: establish a pairwise comparison matrix; step six : Execute the Delphi AHP to obtain the selection result; Step 7: Determine whether the selection result conforms to the screening criteria, output the result if it matches, and execute Step 3 if it does not.

Description

A customizable open source software selection method for different domains

technical field

The invention relates to the technical field of software engineering, in particular to a customizable open source software selection method for different fields.

Background technique

In recent years, with the development of computer and Internet technology, open source software has been successfully applied to important industries related to the national economy and people's livelihood, such as commerce, finance, medical care, e-government, manufacturing, retail, communications, and transportation. In their respective fields, according to different needs, software companies or other organizations have released a large number of powerful open source products. Development engineers can directly use existing software or expand and revise on the basis of open source code to develop products they need . However, the abundance of software types and functions does not mean that users can easily find suitable software. On the contrary, it also increases the difficulty of selection, and using the correct software is the key to the success of the invention. For example, researchers in the field of bioinformatics often use two open source frameworks for natural language processing, UIMA (Unstructured Information Management Architecture) and GATE (General Architecture for Text Engineering), to develop new products. Although there are many similarities between the two, which one is more OK, which one should I choose? How to choose? It has always puzzled researchers in the field of bioinformatics.

Studies have shown that among the many open source products, developers rarely use formal methods to choose the right product, they always choose based on the familiarity of the product or the method recommended by colleagues. High-quality open source software is the premise of realizing high-quality domain software. Therefore, how to use scientific methods to select high-quality products from numerous open source software is one of the difficulties faced by software development engineers in different fields.

SUMMARY OF THE INVENTION

The purpose of the present invention is to propose a customizable open source software selection method for different fields in view of the lack of a customizable open source software selection method for different fields in the prior art.

The technical solution adopted by the present invention in order to solve the above-mentioned technical problems is: a customizable open source software selection method for different fields, comprising the following steps:

Step 1: Customize metrics according to the application field of the software, and then collect metrics data;

Step 2: Select a supervisory group;

Step 3: Publish the collected metrics data information to the crowdsourcing workers;

Step 4: Crowdsourcing workers determine the weights of the first and second layers of the quality model and score them. The supervision team will screen the obtained results and determine whether they meet the screening criteria. If they meet the requirements, go to Step five. three;

Step 5: Establish a pairwise comparison matrix;

Step 6: Execute the Delphi AHP to obtain the selection result;

Step 7: Determine whether the selection result complies with the screening criteria, and output the result if it does. If it does not, go to Step 3.

Further, the screening criterion is N×M×P, where N is the number of crowdsourcing workers, M is the number of feedbacks obtained, and P is the field to which the crowdsourcing workers belong, N≥M, M≥3P, P> 1.

Further, the crowdsourcing worker adopts a recommendation mode.

Further, the Delphi AHP is replaced with crowdsourcing technology, and the specific steps are as follows: replace the domain experts in the Delphi AHP with crowdsourcing workers in different fields, and then, according to the three decision-making ideas, from three The feedback results of the crowdsourcing workers are screened in each aspect, that is, the number of crowdsourcing workers, the number of feedback received and the fields of the crowdsourcing workers.

Further, the metric adopts the ISO/IEC9126 quality model.

Further, the software also includes new quality features included in the ISO/IEC9126 quality model.

Further, the steps of including the new quality feature in the ISO/IEC9126 quality model are: firstly, the new quality feature is compared with the quality sub-features in the ISO/IEC9126 quality model, and then the sub-features not involved in the original model are compared. Added below the corresponding mass feature in the second layer of the model.

The beneficial effects of the present invention are as follows: the present invention firstly adopts the crowdsourcing technology, and replaces the domain experts in the Delphi analytic hierarchy process with crowdsourcing workers in different fields, and then, according to the three decision-making ideas, the crowdsourcing work is analyzed from three aspects. According to the feedback results of the users, the number of crowdsourcing workers, the number of feedback received and the field of the crowdsourcing workers are screened, so as to give an objective evaluation of the software quality and help people choose the most suitable software.

Description of drawings

FIG. 1 is an overall flow chart of the present invention.

Detailed ways

Embodiment 1: This embodiment is described in detail with reference to the figure. In this embodiment, a customizable open source software selection method for different fields includes the following steps:

Step 1: Customize metrics according to the application field of the software, and then collect metrics data;

Step 2: Select a supervisory group;

Step 3: Publish the collected metrics data information to the crowdsourcing workers;

Step 4: Crowdsourcing workers determine the weights of the first and second layers of the quality model and score them. The supervision team will screen the obtained results and determine whether they meet the screening criteria. If they meet the requirements, go to Step five. three;

Step 5: Establish a pairwise comparison matrix;

Step 6: Execute the Delphi AHP to obtain the selection result;

Step 7: Determine whether the selection result complies with the screening criteria, and output the result if it does. If it does not, go to Step 3.

The present invention firstly adopts crowdsourcing technology to replace domain experts in Delphi analytic hierarchy process with crowdsourcing workers in different fields, and then, according to three decision-making ideas, the feedback results of crowdsourcing workers are screened from three aspects, That is, the number of crowdsourcing workers, the number of feedback received, and the field of crowdsourcing workers. Based on this idea, the present invention provides the following definitions.

Define the screening criteria for crowdsourcing feedback results: N×M×P. N: the number of crowdsourcing workers; M: the number of feedbacks obtained; P: the field to which the crowdsourcing workers belong, where N≥M, M≥3P, P>1.

For software products, people in different roles have different concerns. Users, product developers, and managers all pay attention to product quality from their own perspectives. Product quality should be the comprehensive expression of these points of view, but no software can satisfy all interests at the same time. stakeholder needs. Therefore, using crowdsourcing technology and iterative feedback from crowdsourcing workers in different fields can help to give an objective evaluation of software quality, and finally get the selection result.

Because the research object of the present invention is a cross-domain software product, in the feedback results, the crowdsourcing workers who provide feedback should be from related fields, so as to ensure that feedback information can be obtained in each field. Sometimes, the feedback information provided by different people may be the same, so the number of feedback information obtained (M) is less than or equal to the number of crowdworkers providing feedback (N). In addition, due to the defects of crowdsourcing technology, the feedback results of crowdsourcing workers may have invalid information. Therefore, according to the three-way decision-making idea, the amount of feedback information (M) obtained is at least 3 times the number of fields (P), thus Ensure the validity and objectivity of the final evaluation results.

On the basis of the above definition, the present invention uses the ISO/IEC9126 quality model as a basis, and proposes a method combining crowdsourcing technology, three-way decision-making and Delphi analytic hierarchy process to realize a field-oriented customizable open source software selection method. The method is shown in Figure 1.

The basic idea of the present invention is: firstly, by comparing and studying the existing software quality models, the present invention selects the ISO/IEC9126 quality model as the basis, and then determines its quality according to the characteristics of software products in specific application fields and the results of relative dependence theoretical evolution analysis sub-features, and compare these mass sub-features with the mass sub-features in the ISO/IEC9126 quality model, and add sub-features not involved in the original model to the second level (due to the comprehensive nature of the quality features in the first layer of the model and extensiveness, so in general new quality features are added in the second layer of the model), enabling domain customizability. Then, collect data on the selected products according to the metrics, and publish the data and other relevant information to the crowdsourcing workers. Among them, the selection of crowdsourcing workers adopts the recommendation (push) model, that is, the crowd testing platform recommends tasks to a group of workers, and they decide whether to accept the tasks according to their own conditions. Then, the relevant weights of the quality features of the first and second layers of the quality model are given and scored by the crowdsourcing workers, and the supervision team will screen the obtained feedback results, and the screening criteria are as described in the definition. If the obtained result set complies with the screening criteria, a paired comparison matrix is established, and the analytic hierarchy process is performed to obtain the selection result. If the obtained result set does not meet the screening criteria, you need to return to Step 3 for crowdsourcing evaluation again. The dotted line in the figure indicates that the evaluation process can be repeated (Step7→Step3), so that the supervision group can repeat the evaluation to verify the selection result.

Example verification: The present invention takes two open source software frameworks UIMA and GATE in the field of biological information as examples to describe the execution process of the method in detail.

First, determine the quality characteristics of software according to the characteristics of software products in the field of bioinformatics and the results of relative dependence theoretical evolution analysis, if some new characteristics are not included in the ISO/IEC9126 model (due to the comprehensive quality characteristics of the first layer of the model and extensiveness, so in general new quality features are added in the second layer of the model), these new quality features are compared with the quality sub-features in the ISO/IEC9126 quality model, and the sub-features not involved in the original model are compared. The features are added below the corresponding quality features in the second layer of the model. For example, the customizability of software is a quality feature that experts in the field of bioinformatics are more concerned about (customizability refers to the ability of software to expand according to user needs), but the ISO/IEC9126 quality model does not include this feature, so customizability should be is added under the Usability quality feature on the first layer of the model. Then, according to the software metrics and the ISO/IEC9126 quality model, the present invention proposes to list 18 metrics as shown in Table 5. Then, according to these indicators, collect data on the software products to be selected, and publish the data and other relevant information to the crowdsourcing workers in the recommended mode. The crowdsourcing workers should come from the field of software engineering and the field of bioinformatics. Next, the relevant weights of the quality features of the first and second layers of the quality model are given and scored by the crowdsourcing workers, and the supervision team screens the obtained feedback results. The screening criteria are as described in the definition. If the obtained result set meets the screening criteria, a paired comparison matrix is established, and then the AHP is performed to obtain the selection result. If the obtained result set does not meet the screening criteria, you need to return to Step 3 for crowdsourcing evaluation again. The dotted line in the figure indicates that this evaluation process can be repeated, so that the supervision group can repeat the evaluation to verify the selection results.

Table 1 Relationship between quality characteristics and metrics

It should be noted that the specific embodiments are only explanations and descriptions of the technical solutions of the present invention, and cannot be used to limit the protection scope of the rights. Any changes made according to the claims and description of the present invention are only partial changes, which should still fall within the protection scope of the present invention.

Claims (7)

1. A customizable open source software selection method for different fields, characterized by comprising the following steps: Step 1: Customize metrics according to the application field of the software, and then collect metrics data; Step 2: Select a supervisory group; Step 3: Publish the collected metrics data information to the crowdsourcing workers; Step 4: Crowdsourcing workers determine the weights of the first and second layers of the quality model and score them. The supervision team will screen the obtained results and determine whether they meet the screening criteria. If they meet the requirements, go to Step five. three; Step 5: Establish a pairwise comparison matrix; Step 6: Execute the Delphi AHP to obtain the selection result; Step 7: Determine whether the selection result complies with the screening criteria, and output the result if it does. If it does not, go to Step 3. 2. The customizable open source software selection method oriented to different fields according to claim 1, characterized in that: the screening criterion is N×M×P, wherein N is the number of crowdsourcing workers, and M is the number of feedbacks obtained , P is the field of crowdsourcing workers, N≥M, M≥3P, P>1. 3 . The customizable open source software selection method for different fields according to claim 1 , wherein the crowdsourcing worker adopts a recommendation mode. 4 . 4. the customizable open source software selection method oriented to different fields according to claim 1, is characterized in that: described Delphi AHP is replaced with crowdsourcing technology, and its concrete steps are as follows: with crowdsourcing workers in different fields Instead of domain experts in the Delphi analytic hierarchy process, then, according to the three decision-making ideas, the feedback results of the crowdsourcing workers are screened from three aspects, namely the number of crowdsourcing workers, the number of feedback received and the membership of the crowdsourcing workers. field. 5 . The customizable open source software selection method for different fields according to claim 1 , wherein the metric adopts the ISO/IEC9126 quality model. 6 . 6 . The method for selecting customizable open source software for different fields according to claim 5 , wherein the software further includes new quality features, and the new quality features are included in the ISO/IEC9126 quality model. 7 . 7. The customizable open source software selection method oriented to different fields according to claim 6, characterized in that: the step that the new quality feature is included in the ISO/IEC9126 quality model is: firstly compare the new quality feature with the ISO/IEC9126 quality model. The quality sub-features in the IEC9126 quality model are compared, and then the sub-features not involved in the original model are added under the corresponding quality features in the second layer of the model.