阅读说明：本技术 一种软件开发需求质量的评价方法及系统 (Method and system for evaluating software development requirement quality ) 是由 李寒霜 王斌 孟贺 于 2021-07-21 设计创作，主要内容包括：本发明公开了一种软件开发需求质量的评价方法及系统,所述评价方法包括基于测试需求文档的特征项与质量评分之间的线性回归方程,求解得到每一特征项的权值；将待评价需求文档中的特征项与样本需求文档中的特征项进行匹配；将匹配成功的特征项的权值累计求和,得到待评价需求文档的质量得分。本发明以自动评价为主,人工评价为辅,不仅可以提高文档质量审查的效率,节省人力成本和时间,还为人工审查的重点指明了方向。(The invention discloses an evaluation method and a system for software development demand quality, wherein the evaluation method comprises the steps of solving to obtain the weight of each feature item based on a linear regression equation between the feature item and the quality score of a test demand document; matching the characteristic items in the demand document to be evaluated with the characteristic items in the sample demand document; and accumulating and summing the weights of the successfully matched feature items to obtain the quality score of the document to be evaluated. The method takes automatic evaluation as a main part and manual evaluation as an auxiliary part, can improve the efficiency of document quality examination, saves labor cost and time, and indicates directions for the key points of manual examination.)

1. A method for evaluating the quality of software development requirements is characterized by comprising the following steps:

solving to obtain the weight of each feature item based on a linear regression equation between the feature item of the test demand document and the quality score;

matching the characteristic items in the demand document to be evaluated with the characteristic items in the sample demand document;

and accumulating and summing the weights of the successfully matched feature items to obtain the quality score of the document to be evaluated.

2. The method for evaluating the quality of software development requirements according to claim 1, further comprising:

in the process of matching the feature items in the required document to be evaluated with the feature items in the sample required document, if a certain feature item of the required document to be evaluated is found to be missing, full-text field search is conducted on the required document to be evaluated, if the feature item can be searched, the location of the feature item is returned, and after manual review, the feature item is judged to be the feature item which is successfully matched.

3. The method for evaluating the software development requirement quality according to claim 1, wherein the sample requirement document is generated by the following method:

determining a chapter distribution structure of the software requirement specification;

defining chapter titles of the document as characteristic items;

and (4) properly modifying and cutting the characteristic items of the software requirement specification to finally form a sample requirement document.

4. The method for evaluating the software development demand quality according to claim 1, wherein the weight determination method of the feature item comprises the following steps:

randomly selecting N software requirement specification specifications with uneven quality as test documents, wherein N is greater than or equal to the number of characteristic items of sample requirement documents;

respectively extracting feature items of all test documents, and recording the manual review quality score of each test document;

taking the feature items of the test documents as independent variables, taking the document scores as dependent variables, and constructing a linear regression equation between the dependent variables and the independent variables to obtain N independent equation expressions between the feature items and the quality scores;

and solving N independent equation expressions to obtain the coefficient of each characteristic item, namely defining the coefficient as the weight of the characteristic item.

5. The method for evaluating the quality of software development requirements according to claim 1, wherein: the generalized matching process is as follows:

aiming at the characteristic items in the sample requirement document, constructing a characteristic item set of each characteristic item;

and matching the characteristic items in the document to be evaluated with the corresponding characteristic item set.

6. A system for evaluating the quality of software development requirements is characterized by comprising:

the characteristic item weight calculation module is used for solving and obtaining the weight of each characteristic item based on a linear regression equation between the characteristic item of the test demand document and the quality score;

the characteristic item matching module is used for matching the characteristic items in the required document to be evaluated with the characteristic items in the sample required document;

and the scoring module is used for accumulating and summing the weights of the successfully matched feature items to obtain the quality score of the document to be evaluated.

7. The system according to claim 6, wherein the system comprises: the evaluation system further comprises a field search module, wherein the field search module is used for carrying out full-text field search on the document to be evaluated if a certain feature item of the document to be evaluated is found to be missing in the process of matching the feature item in the document to be evaluated with the feature item in the sample requirement document, returning the location of the feature item if the feature item can be searched, and judging the feature item to be a successfully matched feature item after manual review.

8. The system according to claim 6, wherein the system comprises: the sample requirement document is generated by the following method:

determining a chapter distribution structure of the software requirement specification;

defining chapter titles of the document as characteristic items;

and (4) properly modifying and cutting the characteristic items of the software requirement specification to finally form a sample requirement document.

9. The system according to claim 6, wherein the system comprises: the weight determination method of the feature item comprises the following steps:

randomly selecting N software requirement specification specifications with uneven quality as test documents, wherein N is greater than or equal to the number of characteristic items of sample requirement documents;

respectively extracting feature items of all test documents, and recording the manual review quality score of each test document;

taking the feature items of the test documents as independent variables, taking the document scores as dependent variables, and constructing a linear regression equation between the dependent variables and the independent variables to obtain N independent equation expressions between the feature items and the quality scores;

and solving N independent equation expressions to obtain the coefficient of each characteristic item, namely defining the coefficient as the weight of the characteristic item.

10. The system according to claim 6, wherein the system comprises: the generalized matching process is as follows:

aiming at the characteristic items in the sample requirement document, constructing a characteristic item set of each characteristic item;

and matching the characteristic items in the document to be evaluated with the corresponding characteristic item set.

Technical Field

The invention relates to the field of software engineering, in particular to a method and a system for evaluating software development requirement quality.

Background

In the software development process, the requirement analysis is used as a preliminary investigation phase in the early stage, which is the premise of software development and maintenance and directly determines the success or failure of a software project. If the previous requirement analysis work is not complete enough or deviates from the user requirement, the serious consequences can be brought to the whole flow of software development, and the later the requirement problem is found, the higher the repair cost is, and the earlier the problem is found, the lower the repair cost is. Therefore, the intermediate product of the demand analysis must be examined and evaluated at the early stage of software development, so that the existing demand problem can be discovered and corrected at an early stage, thereby improving the efficiency of software development and saving the cost of software maintenance.

In the requirement analysis stage, the intermediate product generated by the development organization is the software requirement specification. At present, the evaluation of the software requirement specification is usually completed by manual detection, so that the working efficiency is low; and no quantifiable evaluation method exists, and the judgment criteria of the document quality are not uniform. Therefore, it is necessary to establish an evaluation method of the requirement specification to quantitatively evaluate the quality of the requirement, so as to better guide the improvement and perfection of the requirement document.

In the chinese patent application with application number CN202010880178.5, a document quality evaluation method is provided for evaluating the design quality of a document to improve the working efficiency, which includes: for each page of the target document, checking a preset checking item; scoring the preset examination items according to a preset scoring standard; and determining the final grade of the document according to the grade result of the preset examination item. Wherein the preset examination items comprise: theme color, theme words, paragraph values, an automatic saving function and an intelligent reloading function. According to the method, the quality of the document is evaluated by examining and grading the theme color, the theme word, the paragraph value, the automatic storage function and the intelligent reloading function of the document, so that the quantitative evaluation of the quality of the document is realized. The method focuses on evaluating the standardization of document writing such as document fonts, paragraph formats, color collocation and the like, neglects the evaluation of the integrity of content items such as document structures, chapter arrangement and the like, and cannot reflect the overall quality of the required text.

In the chinese patent application with application number CN201910968462.5, an evaluation method of a software requirement specification is provided, in which directory entries of the software requirement specification are matched with directory entries with different weights in a pre-stored standard directory one by one, all directory entries in the directory of the software requirement specification are successfully matched, and structural integrity of the requirement document is obtained by calculation. The method matches the catalog of the software requirement specification book with a standard catalog, so as to calculate the structural integrity of the document. In the directory matching process, the method does not consider the diversified expression mode of the same entry in the standard directory, thereby reducing the flexibility and accuracy of directory matching; secondly, the directory matching process excessively depends on the authority of the standard directory, and the analysis of the entry missing reason is omitted; meanwhile, the weight of each directory entry does not give a reasonable determination method, which causes bias of a document evaluation method.

Disclosure of Invention

Aiming at the problems, the invention provides a method and a system for evaluating the software development requirement quality, which take automatic evaluation as the main part and manual evaluation as the auxiliary part, can improve the document quality inspection efficiency, save the labor cost and time, and indicate the direction for the key point of manual inspection.

In order to achieve the technical purpose and achieve the technical effects, the invention is realized by the following technical scheme:

in a first aspect, the present invention provides a method for evaluating quality of software development requirements, including:

solving to obtain the weight of each feature item based on a linear regression equation between the feature item of the test demand document and the quality score;

matching the characteristic items in the demand document to be evaluated with the characteristic items in the sample demand document;

and accumulating and summing the weights of the successfully matched feature items to obtain the quality score of the document to be evaluated.

Optionally, the evaluation method further comprises:

in the process of matching the feature items in the required document to be evaluated with the feature items in the sample required document, if a certain feature item of the required document to be evaluated is found to be missing, full-text field search is conducted on the required document to be evaluated, if the feature item can be searched, the location of the feature item is returned, and after manual review, the feature item is judged to be the feature item which is successfully matched.

Optionally, the sample requirement document is generated by:

determining a chapter distribution structure of the software requirement specification;

defining chapter titles of the document as characteristic items;

and (4) properly modifying and cutting the characteristic items of the software requirement specification to finally form a sample requirement document.

Optionally, the method for determining the weight of the feature item includes:

randomly selecting N software requirement specification specifications with uneven quality as test documents, wherein N is greater than or equal to the number of characteristic items of sample requirement documents;

respectively extracting feature items of all test documents, and recording the manual review quality score of each test document;

taking the feature items of the test documents as independent variables, taking the document scores as dependent variables, and constructing a linear regression equation between the dependent variables and the independent variables to obtain N independent equation expressions between the feature items and the quality scores;

and solving N independent equation expressions to obtain the coefficient of each characteristic item, namely defining the coefficient as the weight of the characteristic item.

Optionally, the process of the generalized matching is:

aiming at the characteristic items in the sample requirement document, constructing a characteristic item set of each characteristic item;

and matching the characteristic items in the document to be evaluated with the corresponding characteristic item set.

In a second aspect, the present invention provides a system for evaluating quality of software development requirements, including:

the characteristic item weight calculation module is used for solving and obtaining the weight of each characteristic item based on a linear regression equation between the characteristic item of the test demand document and the quality score;

the characteristic item matching module is used for matching the characteristic items in the required document to be evaluated with the characteristic items in the sample required document;

and the scoring module is used for accumulating and summing the weights of the successfully matched feature items to obtain the quality score of the document to be evaluated.

Optionally, the evaluation system further includes a field search module, configured to, in a process of matching a feature item in the to-be-evaluated demand document with a feature item in the sample demand document, if a certain feature item of the to-be-evaluated demand document is found to be missing, perform full-text field search on the to-be-evaluated demand document, if the feature item can be searched, return the location of the feature item, and after manual review, determine that the feature item is a feature item that is successfully matched.

Optionally, the sample requirement document is generated by:

determining a chapter distribution structure of the software requirement specification;

defining chapter titles of the document as characteristic items;

and (4) properly modifying and cutting the characteristic items of the software requirement specification to finally form a sample requirement document.

Optionally, the method for determining the weight of the feature item includes:

randomly selecting N software requirement specification specifications with uneven quality as test documents, wherein N is greater than or equal to the number of characteristic items of sample requirement documents;

respectively extracting feature items of all test documents, and recording the manual review quality score of each test document;

taking the feature items of the test documents as independent variables, taking the document scores as dependent variables, and constructing a linear regression equation between the dependent variables and the independent variables to obtain N independent equation expressions between the feature items and the quality scores;

and solving N independent equation expressions to obtain the coefficient of each characteristic item, namely defining the coefficient as the weight of the characteristic item.

Optionally, the process of the generalized matching is:

aiming at the characteristic items in the sample requirement document, constructing a characteristic item set of each characteristic item;

and matching the characteristic items in the document to be evaluated with the corresponding characteristic item set.

Compared with the prior art, the invention has the beneficial effects that:

the invention combines automatic evaluation and manual evaluation, not only improves the efficiency of document review and evaluation, saves labor cost and time, but also can indicate key points and directions for subsequent manual review. Quantitative evaluation of the required quality is realized by grading the quality of the required document to be evaluated, so that the improvement and perfection of the required document can be better guided, and a foundation is laid for the smooth development of the whole software development process.

Drawings

In order that the present disclosure may be more readily and clearly understood, reference is now made to the following detailed description of the present disclosure taken in conjunction with the accompanying drawings, in which:

FIG. 1 is a schematic flow chart of the method for evaluating the quality of demand according to the present invention;

fig. 2 is a schematic structural diagram of the system for evaluating the quality of demand according to the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail with reference to the following embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and do not limit the scope of the invention.

The following detailed description of the principles of the invention is provided in connection with the accompanying drawings.

Software requirements, as the first part of the software lifecycle, are an important bridge between users and software developers. In order to guarantee the effective development of the whole software development process, the quality evaluation needs to be carried out on a main product of a software requirement stage, namely a requirement specification, so that favorable guidance is provided for the subsequent development direction. The method takes automatic evaluation as a main part and manual evaluation as an auxiliary part, can improve the efficiency of document quality examination, saves labor cost and time, and indicates directions for the key points of manual examination.

Example 1

The embodiment of the invention provides a method for evaluating software development demand quality, wherein the demand quality, namely a demand specification, comprises the following steps:

solving to obtain the weight of each feature item based on a linear regression equation between the feature item of the test demand document and the quality score;

matching the characteristic items in the demand document to be evaluated with the characteristic items in the sample demand document;

and accumulating and summing the weights of the successfully matched feature items to obtain the quality score of the document to be evaluated.

In the process of matching the feature items in the required document to be evaluated with the feature items in the sample required document, if a certain feature item of the required document to be evaluated is found to be missing, full-text field search is conducted on the required document to be evaluated, if the feature item can be searched, the location of the feature item is returned, and after manual review, the feature item is judged to be the feature item which is successfully matched.

The sample requirement document is generated by the following method:

determining a chapter distribution structure of the software requirement specification;

defining chapter titles of the document as characteristic items;

and (4) properly modifying and cutting the characteristic items of the software requirement specification to finally form a sample requirement document.

The weight determination method of the feature item comprises the following steps:

randomly selecting N software requirement specification specifications with uneven quality as test documents, wherein N is greater than or equal to the number of characteristic items of sample requirement documents;

respectively extracting feature items of all test documents, and recording the manual review quality score of each test document;

taking the feature items of the test documents as independent variables, taking the document scores as dependent variables, and constructing a linear regression equation between the dependent variables and the independent variables to obtain N independent equation expressions between the feature items and the quality scores;

and solving N independent equation expressions to obtain the coefficient of each characteristic item, namely defining the coefficient as the weight of the characteristic item.

The generalized matching process is as follows:

aiming at the characteristic items in the sample requirement document, constructing a characteristic item set of each characteristic item;

and matching the characteristic items in the document to be evaluated with the corresponding characteristic item set.

In the specific implementation process, the method in the embodiment of the present invention may be implemented according to the following process.

First, section titles (hereinafter, collectively referred to as feature items) of a sample requirement document and a requirement document to be evaluated are extracted, and the sample requirement document and the requirement document to be evaluated are subjected to 'generalized' matching. In the matching process, if a certain feature item of the document to be evaluated is found to be missing, possibly the content of the document is missing or the feature item is covered only due to paragraph merging, the document to be evaluated needs to be subjected to full-text field search. If the feature item can be searched, the location of the feature item is returned, and further manual review is required to confirm the expression integrity of the feature item. And finally accumulating the successfully matched feature item weights to obtain the quality score of the document to be evaluated.

The generation method of the sample requirement document comprises the following steps: and determining the section distribution structure of the software requirement specification by referring to the authoritative national standard GB/T9385 computer software requirement specification. Secondly, according to the experience accumulation of developers and the characteristics of the software to be developed, the feature items of the software requirement specification are modified and cut appropriately, and finally a sample requirement document is formed.

The weight determination method of the feature item comprises the following steps: randomly selecting N (N is more than or equal to the number of the characteristic items of the sample requirement documents) parts to compile requirement specification (hereinafter referred to as test documents) with uneven quality, respectively extracting the characteristic items of all the test documents, and recording the manual review quality score of each test document. Taking the feature item of the test document as an independent variable, taking the document score as a dependent variable, obtaining N independent equation expressions between the feature item and the quality score by constructing a linear regression equation between the dependent variable and the independent variable, and defining the coefficient of each feature item obtained by solving as the weight of the feature item.

The definition of the document feature item 'generalized' matching is as follows: considering that the same characteristic item can have a plurality of different expressions, for example, the characteristic item "term and definition" can be written as "noun explanation", "reference material" can be written as "reference file", etc., by constructing the characteristic item set of each characteristic item, the "generalized" matching is realized, and the flexibility and accuracy of the matching of the characteristic items are improved.

The method for calculating the quality score of the to-be-evaluated required document comprises the following steps: and matching the characteristic items in the to-be-evaluated requirement document with the characteristic items of the sample requirement document, and accumulating and summing the weight values of the successfully-matched characteristic items to obtain the quality score of the to-be-evaluated requirement document. Likewise, the method of calculating the document quality score is not limited thereto, and may be adjusted and improved according to the actual situation.

The necessity of the manual review is as follows: in the process of matching the feature items by the system, if a certain feature item is found to be missing, the system may be caused by a plurality of reasons:

(1) the author considers the content of the feature item in the writing process of the document, so that the content of the feature item is lost;

(2) the author merges the contents of a plurality of chapters in the writing and typesetting process of the document, so that the characteristic items are covered;

(3) the feature item set of each feature item is not comprehensive enough, which results in failure of "generalized" matching of the document feature items.

Thus, further manual review is required to determine if feature terms are truly missing, resulting in a more accurate document quality score.

The evaluation method in the examples of the present invention will be described in detail below with reference to FIG. 1.

a) After the process is started, firstly, initializing a program;

b) selecting and designing a proper sample document, and extracting feature items of the sample document;

c) constructing a feature item set of each feature item;

d) establishing a plurality of linear regression equations between the feature items of the test document and the manual scores, and solving to obtain the weight of each feature item;

e) matching the feature items of the to-be-evaluated requirement document with the sample feature item set;

f) after the feature item matching is finished, calculating the quality score of the document according to different weights of each feature item;

synchronously, if the feature item is found to be missing after matching, the full text needs to be subjected to field search. If the characteristic item can be searched, the position is returned, the follow-up examination is completed manually, and the scoring result of the document is returned.

Example 2

The embodiment of the invention provides a system for evaluating software development demand quality, as shown in fig. 2, comprising:

the characteristic item weight calculation module is used for solving and obtaining the weight of each characteristic item based on a linear regression equation between the characteristic item of the test demand document and the quality score;

the characteristic item acquisition module is used for acquiring characteristic items in the required document to be evaluated and characteristic items in the sample required document;

the characteristic item matching module is used for matching the characteristic items in the required document to be evaluated with the characteristic items in the sample required document;

and the scoring module is used for accumulating and summing the weights of the successfully matched feature items to obtain the quality score of the document to be evaluated.

And the field searching module is used for carrying out full-text field search on the document to be evaluated if a certain characteristic item of the document to be evaluated is found to be missing in the process of matching the characteristic item in the document to be evaluated with the characteristic item in the sample requirement document, returning the positioning of the characteristic item if the characteristic item can be searched, and judging the characteristic item to be a successfully matched characteristic item after manual review.

The sample requirement document is generated by the following method:

determining a chapter distribution structure of the software requirement specification;

defining chapter titles of the document as characteristic items;

and (4) properly modifying and cutting the characteristic items of the software requirement specification to finally form a sample requirement document.

The weight determination method of the feature item comprises the following steps:

randomly selecting N software requirement specification specifications with uneven quality as test documents, wherein N is greater than or equal to the number of characteristic items of sample requirement documents;

respectively extracting feature items of all test documents, and recording the manual review quality score of each test document;

taking the feature items of the test documents as independent variables, taking the document scores as dependent variables, and constructing a linear regression equation between the dependent variables and the independent variables to obtain N independent equation expressions between the feature items and the quality scores;

and solving N independent equation expressions to obtain the coefficient of each characteristic item, namely defining the coefficient as the weight of the characteristic item.

The generalized matching process is as follows:

aiming at the characteristic items in the sample requirement document, constructing a characteristic item set of each characteristic item;

and matching the characteristic items in the document to be evaluated with the corresponding characteristic item set.

In the implementation process, the system in the embodiment of the invention can work according to the following process.

First, section titles (hereinafter, collectively referred to as feature items) of a sample requirement document and a requirement document to be evaluated are extracted, and the sample requirement document and the requirement document to be evaluated are subjected to 'generalized' matching. In the matching process, if a certain feature item of the document to be evaluated is found to be missing, possibly the content of the document is missing or the feature item is covered only due to paragraph merging, the document to be evaluated needs to be subjected to full-text field search. If the feature item can be searched, the location of the feature item is returned, and further manual review is required to confirm the expression integrity of the feature item. And finally accumulating the successfully matched feature item weights to obtain the quality score of the document to be evaluated.

The generation method of the sample requirement document comprises the following steps: and determining the section distribution structure of the software requirement specification by referring to the authoritative national standard GB/T9385 computer software requirement specification. Secondly, according to the experience accumulation of developers and the characteristics of the software to be developed, the feature items of the software requirement specification are modified and cut appropriately, and finally a sample requirement document is formed.

The weight determination method of the feature item comprises the following steps: randomly selecting N (N is more than or equal to the number of the characteristic items of the sample requirement documents) parts to compile requirement specification (hereinafter referred to as test documents) with uneven quality, respectively extracting the characteristic items of all the test documents, and recording the manual review quality score of each test document. Taking the feature item of the test document as an independent variable, taking the document score as a dependent variable, obtaining N independent equation expressions between the feature item and the quality score by constructing a linear regression equation between the dependent variable and the independent variable, and defining the coefficient of each feature item obtained by solving as the weight of the feature item.

The definition of the document feature item 'generalized' matching is as follows: considering that the same characteristic item can have a plurality of different expressions, for example, the characteristic item "term and definition" can be written as "noun explanation", "reference material" can be written as "reference file", etc., by constructing the characteristic item set of each characteristic item, the "generalized" matching is realized, and the flexibility and accuracy of the matching of the characteristic items are improved.

The method for calculating the quality score of the to-be-evaluated required document comprises the following steps: and matching the characteristic items in the to-be-evaluated requirement document with the characteristic items of the sample requirement document, and accumulating and summing the weight values of the successfully-matched characteristic items to obtain the quality score of the to-be-evaluated requirement document. Likewise, the method of calculating the document quality score is not limited thereto, and may be adjusted and improved according to the actual situation.

The necessity of the manual review is as follows: in the process of matching the feature items by the system, if a certain feature item is found to be missing, the system may be caused by a plurality of reasons:

(1) the author considers the content of the feature item in the writing process of the document, so that the content of the feature item is lost;

(2) the author merges the contents of a plurality of chapters in the writing and typesetting process of the document, so that the characteristic items are covered;

(3) the feature item set of each feature item is not comprehensive enough, which results in failure of "generalized" matching of the document feature items.

Thus, further manual review is required to determine if feature terms are truly missing, resulting in a more accurate document quality score.

The foregoing shows and describes the general principles and broad features of the present invention and advantages thereof. It will be understood by those skilled in the art that the present invention is not limited to the embodiments described above, which are described in the specification and illustrated only to illustrate the principle of the present invention, but that various changes and modifications may be made therein without departing from the spirit and scope of the present invention, which fall within the scope of the invention as claimed. The scope of the invention is defined by the appended claims and equivalents thereof.