Software quality and quality model ---- Software quality is a complex concept, and the different people will see the software quality issues from different perspectives. There will be different solutions. From the user's point of view, quality is to meet the needs of customers; from the perspective of developers, quality is consistent with demand instructions; from the point of view, quality is the intrinsic characteristics of the product; from the perspective of value, quality is customer Whether it is willing to buy. ---- In the process of making the software, the quality in the eye of the project is the soft product produced by "full" groundwork to complete the pre-phase function. The "full", including functionality, sex, interface needs and other indicators, such as reliability, sufficiency, reuse and positiveness. However, in practice, once a question, the item managers must be right to emphasize the proficiency, make the house, sacrificed an additional one or several indicators when satisfying a certain indicator. For example, if it is delivered in accordance with period, you need to classify the software function to classify, in the first version, real prior grade functionality, in the second version of the best pre-first level function. Therefore, the item matters need a quality model and metric method for its intelligence. One side of this model can help projects with a software producer that produces a compliant standard, and the other side helps the entry of the risk of the product quality. ---- As mentioned above, the item is required to help him to evaluate the quality of the software and the instructions of the risk. There is a lot of quality models that have been used, and they define different software quality properties. The three quality modes than the more common are McCall model (1977), Boehm model (1978) and ISO 9126 (1993). ---- In the beginning of the project, the project is responsible for the characteristics of the customer, selecting a set of item management objectives, needle to this set of objects, and set a set of data measurements and statistics. This group of managers should be related to the properties of the software product and software, and these attributes can reflect the probability of the target. ---- Example, from the real-purpose angle, we can choose a group of management objectives below: the quality of the demand, the quality of the product, the effectiveness of the development, and the effectiveness of testing. ---- This set of targets reflects both the quality of the software product, and reflects the quality of the software process. The model of the model must be measured by a set of attributes to help define and classify the risk.
Risk ---- The risk is the ability to suffer from loss. Software Development Risk Management Project Includes the Identification, Evaluation and Sampling and Risk of Risk. ---- For large number of software, it is easy to produce the risk of functional quality zone: positive, reliable, sufficiency, reuse and delivery. ---- The order of the above-mentioned wind danger zone can be different from the undesirable items and different items. It is a two-point request for the quality model: First, it is necessary to seek the quantization of the model support to the risk; the second is to seek model support for the risk assessment of the entire project. That is, the measurement of the risk should be usable to the risk of a single piece or the entire project. ---- When the risk of item is known and quantified, they should be classified. If you can grade it in accordance with the different measures: ---- Low level is advanced according to the current trend, it is very likely to achieve its goals, there is no need to set emergency plan; ---- Intermediate According to the current trend, it is possible to achieve its goals, need to formulate Emergency plan; ---- Advanced According to the current trend, it is difficult to achieve its goals, and the emergency plan is required immediately. ---- Classification work is often required to be supported by the data, and should be completed by the experienced analysis. Quality Model and Risk Assessment ---- Once I've selected the management target, and defines the quality attribute of the correlation, it should be determined to measure the measurement. These measured standards are applicable to the entire software life cycle, including the initial stage, which will help the initial phase of the initial issue in the initial period of software. Quality is required ---- 1. Quality and its risk ---- Demand is the basis for the driving of the software, and the mass attribute of the need to be related to the mild determination, completeness, canal, demand, can track. ---- Software needs to define the need for functionality, that is, which work should be done; if the performance needs, how much is to do, how fast it is, the interface needs, that is, how to interface, how to interface, how to interface. If you need to be unclear, it is not easy to solve, and the final product cannot be satisfied with the risk of satisfaction. ----2 . Quality attributes and their measurements ---- (1) Clearness ---- Unclear demand refers to those needs, such demand is realized, different developers will have different understandings. . There are two sets of measurement data to evaluate the clearness of the needs of the document: a group is an unclear demand, a group is an uncertain demand. ---- Unclear, suitable, suitable, not limited to, normal, at least, no time.
---- Uneasite use: can be, also performed as a candidate. ---- Unestability of demand can be reflected in the need to be used in an unclear or uncertain demand. ---- (2) Total ---- The complete demand document requires the description of the needs to be detailed enough to make the design and implementation can be on this basis. The number of measurements such as "to be fixed" or "to be supplemented" can be used to evaluate the number of "to be set" or "to be supplemented" in the selection of documents. ---- (3) Helievability ---- There are two standards that can reflect the rational solution of demand: one is the need to seek the number of documents, and the other is a readability evaluation. Tongmong, the demand, the demand (ie, its deepness and the wideness) reflects the quality of the need to ask the document; and the readability of the index is also an important factor in the impact of the effect. ---- (4) The demand waveability ---- the waveability of the demand refers to the frequency of change in demand. The need to change the impact on the soft material with the addition of the processed work, the larger, and the measurement of the waveability is a hundred points of the total number of changes to the number of changes to the number of requirements in a certain time period. The statistics of the change entries should come from the patriarchal configuration management system. ---- (5) can track ---- Software demand is from the system demand, the system needs should be tracked for it to ensure that the software needs to be in the system set under the system. Software needs to be designed, implemented, and test procedures to ensure that it is the positive identity of a part of the software. There are two kinds of trackable measured standards, and one is not available to the system of software, and the number of software needs to be followed by the system; the other is not to follow the software that is from the code and test. ---- 3. It is necessary to seek risk ---- It has been recognized that the quality of quality is poor and multi-change needs to be the main source of item risk. The demand of waveness is a heavy factor in the risk. It is necessary to measure the amount of force to measure it to reduce its impact. The more you need to change, the greater the influence of the birth, the greater the risk. A value given is that the measurement of the need to seek risk should be constructed throughout the soft part of life. Production quality ---- 1. Production quality and its risk ---- Software discharge project "The main purpose of the project is the code and documentation of the production of the entrants. Therefore, the quality attribute of the measurement required to measure is: system / structural, reuse, maintenance and document and code.
----2 . Software quality attributes and measurement ---- system / structures, reuse and sustainability make the same quality attribute, including complexity, interpretation, and the relationship between the two. ---- (1) complexity ---- people have generally believed that high complex modules are not only difficult to solve, and it is easy to be wrong. Therefore, complexity has a straight impact on the quality of the software and the maintenance and researcher. ---- Measurement of complexity, there are many types of targets, such as: ---- logical complex linear independence test path number; ---- Data complex data type and parameter password; Call the complexity from the module, the number of calls; ---- Nested depth of nested level condition statement. ---- The testability and sundemability of logic complexity and programs are directly related. ---- (2) Program Size ---- The measurement of the program is mainly calculated by the number of lines, the counting method must have: the number of lines, the header, definition, execution, and unable statements), non-annotation line And the number of non-air code lines can be implemented in the number of lines. ---- The size of the module is a label of the mass, and the general work standard is 50 to 100 lines of each module. A larger module will be difficult to solve, and it will reduce sustainability and reuse. ---- 3. Internal document ---- document is the code of the code. It can be a manual that is separated from the code, or it can be an comment in the program. For the measurement of the internal document, we recommend choosing the comments in the program. ---- When calculating the number of comments, the number of online and non-online note should be constructed. The online comments are enclosed in the privileged statement, and many tools ignore the statistics to the online note. Real cost-effectiveness and its risk - - Real cost-effectiveness refers to the effectiveness of the maximum level of resources in the activity of the project. The property and its measurement mark with this goal are: ---- The use of resources of the resource in the unique order of the item. The source measurement is only a single bit by "human", and the measuring common needle is performed on the active movement of the life cycle. These activities are defined by the item. The main activity of the measurement of the measurement is: analysis, code, test, modification, training, etc. ---- The percentage of the percentage of a certain activity or the completion of the task. The yield of the task and product part is a sign of the product. If there is a detailed plan, the completion rate can be measured, and the ratio of the plan and the actual completion situation can be made.
---- It is necessary to strongly adjust that the illegitimate rate of resources and the completion rate of the task should not be used to calculate the production rate of the developer. Once, the developer will go to the cooperation. ---- In the design stage, even in real stages, if a large amount of energy is spent on the activity that needs to be off, the item will face a lot of risk. The effectiveness of testing and its risk ---- The purpose of the effective test is to find and modify the mistakes in the software, know how to make the wrong mistake, press the time to measure the test work, so that the software can make a full work. ---- Once the code is passed through a single-unit test, the official test (including the system test, the integration test and testing test) will be started. The purpose of this process is to find the sub-system and the error between the components from the unpredictable interaction, and independently, the system is provided to provide the functions described in the need. ---- Test effectiveness and measurement is: ---- Positiveness is here, the correct definition is that the code is completed into need to define the degree of defining, and the one is that the software must have nothing wrong. The wrong mistake is found in the test and reciprocity. ---- The correct measurement of the standard is the wrong message: check the date, repair the completion date, the error, the test case number, the wrong source, the modified code is changed. ---- Through the statogram, the misunderstandings found in the test were in the case of the measuring point of the meter, which was as follows: --- 即 Time to the time, the error of the error will reach a limit. ---- Positiveness is the number of misunderstandings that reported the presentation. When the test process proceeds to 1/3, according to the test curve, the total error number is predicted. ---- The number of error errors that pass the code module and the subsystem can be pre-measured. That is, the code segment of the generated error number exceeds average the average value should be paid, and whether it should be returned. Joint ventilation estimates ---- Estimation of some risk, also need to consider the measurement results of other quality objectives. For example, the high demand wave movement is only the number of variations that are not measured by the demand, and it must also be subject to the time of the need to seek active activity after the demand stage end bundle. A very meaningful measurement is the average time to spend every need to change, which is the total time in which the total demand variation is taken in addition to the time period. ---- Another existing risk assessment measurement material is the number of key erroneous erroneous erroneous errors in the module in the module in the same time.
