【正文】
native, sales slips might be collected throughout the day and processed as a batch after the store closes. In the first system, the two programs would deal with direct access files。當工資程序執(zhí)行時,程序存取數據 文件。用戶和程序員似乎常常講不同的語言,導致了兩者之間的聯系障礙。理想上,希望達到一個平衡,貨物即充足但又不太多。 目的是確保用戶和分析員兩者都思考同一件事。用戶清楚必須做什么。啟動這項工作的一個好方法是描述功能之間的數據如何流動.顧名思義,為了用圖解方法描述這些數據流,那么數據流程圖就是特別有用的.圖 1中用到四種符號數據的像和目的地用方框表示,輸入的數據從源進入系統(tǒng),而輸出的數據流到目的地。同時, MANAGEMENT(經營部門)可通過 COMMUNICATE(通信)存取系統(tǒng),檢查 STOCK 中的數據,如果需要的話,請求重新訂貨.訂貨單一經核準。 另外一些數據,例如購買的貨物品種和數量由 CUSTOMER 產生,還有一些數據,例如銷售價格和訂貨點由 MANAGEMENT 產生,或者由它提出要求。用戶只是從功能和數據出發(fā)考慮系統(tǒng),他們并不具體 設計程序、文件和硬件,并且在這個初始的帶有決定性的分析階段,分析員必須像用戶而不像程序設計員一樣去思考問題。 13 設計 當進入設計階段的時候,已知道必須做什么事。經營部門用該程序處理數據和核對訂貨單。例如,在某些應用中數據是在有規(guī)律的預定時間內被處理,一般的是整段時間地收集數據,成批地一起處 理。貨物清單文 件應按順序組織還是直接組織呢?面對著這種選擇,一個好的系統(tǒng)分析員應考慮兩種方案。在實現階段,每個程序都用應用軟件技術來編排15 和書寫;要建立文件并檢查文件的內容:購買、安裝和測試新的硬件;另外還要寫出操作步驟,并對系統(tǒng)做出評價。良好的計劃、實用的文件資料和合適的程序結構都有助于減少維護費用。 另外一方面 , 還要 感謝北京科學技術出版社的各位老師們,在課題的調研工作和系統(tǒng)的制作過程 中,科學技術出版社的領導和各位老師也給予我很大的支持,這也是本課題能夠順利地進行的一個重要原因。 感謝印刷學院的老師們,是您們給予我豐富的理論知識,使我開闊了眼界,同時也是您們教育我奮發(fā)圖強。維護延續(xù)在系統(tǒng)的整個生命期內。 實現 系統(tǒng)的主要部分分析清楚了,就能開始開發(fā)這幾 部分。一般來說,這些事務處理系統(tǒng)要求直接存取文件。以貨物清單文件為例,它包含來自 STOCK的所有數據元素,這些數據元素被列在數據字典中.利用這些數據元素能安排文件的數據結構。 系統(tǒng)流程圖用符號表示程 序、過程、硬設備及物理系統(tǒng)中的其它部件(圖3),流程圖(圖 4)表明,通過終端進入系統(tǒng)的事務數據被數據采集程序處理,然后存放在貨單文件上。也許,數據流程圖和數據字典為分析員理解系統(tǒng)的需求提供了最重要的文件,通過查閱這些文件,用戶能改正誤解或疏忽大意。記住,它的目標是把用戶的需求轉換成技術術語。例如,重訂貨單送給了供貨商們,商店可能想要 25 條工作褲,為了填寫訂貨單,供應商需要商品 的說明和再訂貨的數量,那么,這些數據從什么地方來呢?由于數據是 Generate reorder 輸出的,所以既要有數據輸入給它,還要有數據經過它的處理而產生。從 CUSTOMER(顧客)開始,由此事務進入系統(tǒng),在這所辦理的手續(xù)由 PROCESS TRANSATION(事務處理程序)處理。最后,經過經營管理部門的 核準,系統(tǒng)應該產生一個重新訂貨單,準備發(fā)送給供應商。有了好的問題定義和肯定的可行性研究,分析員就能著手計劃和研究問題的解了。用戶幾乎都不是計算機方面的專家,他們大多數人把計算機看成魔術盒,并不關心它是如何工作的。 設想一個小服裝店, 成批地購買貨物,把貨物擺在貨架上,零售給顧客。計算機技術方面的專業(yè)人員,例如程序設計員有這方面的專門知識,但是可能在用戶的專業(yè)領域內缺少訓練。 系統(tǒng)就是為實現一個目標而共同工作的一組部件。s data structure can be planned, How should the file be anized? That depends on how it will be accessed. For example, in some applications, data are processed at regular, predictable intervals. Typically, the data are collected over time and processed together, as a batch. If batch processing is acceptable, a sequential file anization is probably best. It is not always possible to wait until a batch of transactions is collected, however. For example, consider an air defense early warning system. If an unidentified aircraft is spotted it must be identified immediately the idea of waiting until 5 _ 00 . because that39。s understanding of the system requirements. By reviewing these documents, the user can correct misunderstandings or oversights. Finally, they represent an excellent starting point the next step, design. 23 Design As we enter the design stage, we know what the system must do, and thus can begin thinking about how to do it. The objective is to develop a strategy for solving the problem. At this stage, we are not interested in writing code or in defining precise data structures。 input data enter the system from a source, and output data flow to a destination. Once in the system, the data are manipulated or change by processes, represented by roundcorner rectangles. A process might be a program, a procedure, or anything else that changes or moves data. Data can be held for later processing in data stores, symbolized by openended rectangles. A data store might be a disk file, a tape file, a database, written notes, or even a person39。s owner) needs. Note that, as the process begins, the user possesses the critical information, and the analyst must listen and learn. Few users are technical experts. Most see the puter as a magic box, and are not concerned with how it works. At this stage, the analyst has no business even thinking about programs, files, and puter hardware, but must municate with the user on his or her own term. The idea is to ensure that both the user and the analyst are thinking about the same thingThus, a clear, written statement expressing the analyst39。t enough。 this knowledge is extracted and formally documented. Most users think in terms of the functions to be performed and the data elements to be manipulated. The objective is to identify and link these key functions and data elements, yielding a logical system design. Start with the sys