The third of reading notes of demand analysis and system design

The third of reading notes of demand analysis and system design

I summarize the knowledge content from the eighth chapter "Database Design": Class model andbcedThe class package reflects the application class, not the database structure, the entity class represents the persistent database object in the application, but not the permanent class in the database; A persistent database layer can be a relational database, an object relational database, or an object database; A database model is an abstraction that represents the structure of a database, containing three abstractions. are: External data model, logical data model and physical data model; Database packages are not able to export database models, which are derived from database modeling. The mapping between the application and the database, which is responsible for the database package, can be a complex problem, with two difficulties in mapping: the storage structure of the database does not do anything to the object-oriented paradigm, and secondly the database is almost not designed for a single application. The main purpose of object database management system is to make transparent integration between database and application programming language, and its modeling language is object and text, each object has aOID, the text does notOID, whose value is its identifier;ODBone of the main benefits is the internal support for text and object types, which makesODBbecome object-oriented isdevelopment of the natural realization Platform,it supports correlation and generalization relationships (aggregations are supported by forcing associations);ODMGThe object model defines two kinds of generalization relationships:ISA(equivalent to inheritance of interfaces) andEXTENDSrelationship (equivalent to implementing inheritance);Odbmswith internal operations to support text and object types;ODBtask is to establish an object-oriented implementation for the object-oriented model, which can map entity classes, associations, aggregations, and generalizations. The Object relational database combines the old-fashioned relational model with the modern object model, where the columns in the table can take an internal or user-defined type as its value; The Object table is a collection of columns with one or more columns; row types allow tables to have relatively complex internal structures without even needing to use structured types or collections ; structured types can be used to define reference types; mappings are not for abstractSQLStandard to do, but for the actualORDBto do;RDBThe main modeling language in the model is a relational table of columns, which does not support object types, structured types, collections, and references; relational databases define data with a column or row table, and a relational table is defined by a fixed set of columns ;RDBThe model uses referential integrity constraints to maintain relationships between tables, and rules and descriptive referential integrity constraints support the definition of simple business rules in a database; a relational view is a stored and namedSQLquery; relate toRDBmapping involves using referential integrity constraints between tables, and triggers can be used to suppress theUMLThe semantics that are captured in the business rules in the class model are programmed, and normalization will further affect this mapping.

      from the Nineth chapter, "Program and transaction Design", learned the following: well-designed programs to maximize the cohesion of a class and to minimize the coupling of the class; demeter c/s sql Five levels of the interface , fifth level sql allow users to program the database directly Stored procedures and triggers severely affect the server aspect of the program design, which extends the window navigation diagram to include database considerations. A transaction is a logical unit of database work, which begins with a consistent state of the database and ensures that it is at the end of a consistent condition, that the transaction guarantees database concurrency and database recovery, that traditional database applications require short transactions, and that some new database applications work with long transactions. Two-way engineering is a process in which the design model and the program are synchronized, but they can evolve each other; two-way engineering combines forward engineering from design model and reverse engineering from program; bidirectional engineering is generally independent of client programs and database programs.

The tenth chapter of "Test and Change Management" describes the following main content: The test is divided into system service test and System constraint test two; system service tests can be based on non-execution or execution-based tests, including walk-through and check-based tests, and execution-oriented testing can be tests for specifications or for code. System-constrained testing consists of a number of completely different tests that relate to the user interface, database, authorization, performance, stress, and failure. Configuration and installation, etc., it is carried out in parallel with the system service test, while others are done independently. Test and change management requires specific documentation such as test plans, test case documentation, failure and improvement project documentation, and test requirements are identified in the test document and linked to use case requirements in the use case documentation. Change requests are generally either faulty or improved; a change management tool is used to change the submission of a request and save its trajectory, just as the developer treats it; a key part of the change management tool is related to establishing a traceable path between change requests and other system artifacts, especially test requirements and use-case requirements.

The third of reading notes of demand analysis and system design