分析模式

前言

我們幾個(gè)所謂的“四人幫”在寫《設(shè)計(jì)模式》時(shí)就知道，除了面向?qū)ο笤O(shè)計(jì)模式以外，還有許多其他軟件模式。到寫完書后，我們已經(jīng)看到了分布式編程模式、用戶界面模式，甚至是組織軟件開發(fā)團(tuán)隊(duì)的模式。但我們還沒見過很清晰的面向?qū)ο蠓治瞿Ｊ健eter Coad提出的模式是最接近于面向?qū)ο蠓治瞿Ｊ降?，但這些模式和我們的模式非常像。而在我們看來(lái)，純分析模式應(yīng)該是大不一樣的。在讀Martin Fowler的書稿時(shí)，我終于找到了我一直夢(mèng)寐以求的東西。他講的模式容納了大量領(lǐng)域知識(shí)，又能在所有商業(yè)軟件中都適用。和設(shè)計(jì)模式一樣，這些模式很抽象，足以幫助你的軟件應(yīng)對(duì)各種猝不及防的需求變化，但又非常具體，容易理解。它們不見得是最明顯的解決建模問題的方法，但在我看來(lái)真實(shí)可信。我以前見過許多這類解決方法，都很有效。與其說(shuō)我是一個(gè)建模師，不如說(shuō)我是個(gè)設(shè)計(jì)師。在Martin Fowler描述的很多領(lǐng)域，我都沒有太多經(jīng)驗(yàn)。雖然我覺得那些模式很好，但我對(duì)自己的感覺不是太自信。然而，自從讀了這本書之后，我就開始在軟件項(xiàng)目中嘗試這些模式，并用于教學(xué)。它們的確行之有效！等我看過David Hay寫的Data. Model Patterns一書后，我的自信心大大增強(qiáng)了。我意識(shí)到，盡管他們有不同的背景和專業(yè)用詞，但他們看到了許多相同的模式。模式應(yīng)該描述現(xiàn)實(shí)，而不是發(fā)明新事物，Martin Fowler精確地描述了商業(yè)軟件面向?qū)ο竽Ｐ椭械哪Ｊ健Ｄ愦罂蓪?duì)他描繪的模式滿懷信心。雖然Martin介紹了許多建模原則，但這本書不是一本專講原理的書，你不必非要讀完全書并做相應(yīng)練習(xí)之后才能有所收獲。書中全都是實(shí)用的模式，可以隨學(xué)隨用。針對(duì)遇到的具體問題，找到適用的章節(jié)，你將發(fā)現(xiàn)許多大有助益的好主意。你也可以逐章閱讀，每一章都有新思想。為了充分利用本書，需要知道兩件事。首先，許多模式都深藏不露，比表面看上去要強(qiáng)大得多。比如，.Accountability模式就可以應(yīng)用于幾乎任何項(xiàng)目中。不要局限于表面上明顯適用于你的項(xiàng)目的那幾章，而應(yīng)盡可能多學(xué)習(xí)幾種模式，并且看看是否能派上用場(chǎng)。其次，要保證讓合作者也閱讀本書。

內(nèi)容概要

本書講述各種分析模式和輔助模式，專注于面向?qū)ο蠓治雠c設(shè)計(jì)的結(jié)果——模型本身，給出了來(lái)自金融貿(mào)易、測(cè)量、財(cái)務(wù)以及組織關(guān)系等多個(gè)領(lǐng)域內(nèi)的一系列模式。書中每個(gè)模式都包含了設(shè)計(jì)背后的原理、使用的規(guī)則以及實(shí)現(xiàn)的技巧，給出的例子包含了有用模型的細(xì)節(jié)，并介紹了用于提高分析、建模和實(shí)現(xiàn)的重用技巧。　　本書適用于面向?qū)ο蟮挠?jì)算機(jī)分析人員和設(shè)計(jì)人員、數(shù)據(jù)建模人員、編程人員以及軟件工程師。

作者簡(jiǎn)介

Martin Fowler出生于英國(guó)沃爾索爾?，F(xiàn)定居美國(guó)梅爾羅斯。世界級(jí)軟件開發(fā)大師。OOAD、UML、模式、敏捷開發(fā)方面的領(lǐng)軍人物。自上世紀(jì)80年代中期一直從事面向?qū)ο蠹夹g(shù)的研究。作為獨(dú)立顧問從事咨詢工作十余年。1999年加盟Thought Works，現(xiàn)為該公司首席科學(xué)家。著有《重構(gòu)：改善既有代碼的設(shè)計(jì)》、《企業(yè)應(yīng)用架構(gòu)模式》、《UML精粹》等一系列膾炙人口的經(jīng)典。

書籍目錄

Chapter 1 Introduction　1.1 Conceptual Models　1.2 The World of Patterns 　1.3 The Patterns in this Book　1.4 Conceptual Models and Business Process Reengineering　1.5 Patterns and Frameworks　1.6 Using the Patterns 　ReferencesPart 1. Analysis Patterns　Chapter 2 Accountability 　2.1 Party 　2.2 Organization Hierarchies　2.3 Organization Structure 　2.4 Accountability 　2.5 Accountability Knowledge Level 　2.6 Party Type Generalizations　2.7 Hierarchic Accountability 　2.8 Operating Scopes 　2.9 Post 　References 　Chapter 3. Observations and Measurements　3.1 Quantity 　3.2 Conversion Ratio 　3.3 Compound Units　3.4 Measurement 　3.5 Observation 　3.6 Subtyping Observation Concepts　3.7 Protocol 　3.8 Dual Time Record 　3.9 Rejected Observation 　3.10 Active Observation, Hypothesis, and Projection　3.11 Associated Observation　3.12 Process of Observation　References 　Chapter 4. Observations for Corporate Finance　4.1 Enterprise Segment　4.2 Measurement Protocol 　4.3 Range　4.4 Phenomenon with Range 　4.5 Using the Resulting Framework　References　Chapter 5. Referring to Objects 　5.1 Name 　5.2 Identification Scheme　5.3 Object Merge　5.4 Object Equivalence　References 　Chapter 6. Inventory and Accounting 　6.1 Account 　6.2 Transactions 　6.3 Summary Account 　6.4 Memo Account 　6.5 Posting Rules 　6.6 Individual Instance Method 　6.7 Posting Rule Execution　6.8 Posting Rules for Many Accounts 　6.9 Choosing Entries 　6.10 Accounting Practice 　6.11 Sources of an Entry 　6.12 Balance Sheet and Income Statement 　6.13 Corresponding Account 　6.14 Specialized Account Mode] 　6.15 Booking Entries to Multiple Accounts 　Further Reading 　References 　Chapter 7. Using the Accounting Models　7.1 Structural Models　7.2 Implementing the Structure 　7.3 Setting Up New Phone Services　7.4 Setting Up Calls　7.5 Implementing Account-based Firing　7.6 Separating Calls into Day and Evening 　7.7 Charging for Time　7.8 Calculating the Tax 　7.9 Concluding Thoughts　References 　Chapter 8. Planning 　8.1 Proposed and Implemented Action 　8.2 Completed and Abandoned Actions 　8.3 Suspension 　8.4 Plan 　8.5 Protocol　8.6 Resource Allocation 　8.7 Outcome and Start Functions 　References 　Chapter 9. Trading 　9.1 Contract 　9.2 Portfolio 　9.3 Quote 　9.4 Scenario 　References 　Chapter 10. Derivative ContractsChapter 11. Trading PackagesChapter 12. Layered Architecture for Information SystemsChapter 13. Application FacadesChapter 14. Patterns for Type Model　　Chapter 15. Association Patterns Chapter 16. Afterword Part 3. Appendix

章節(jié)摘錄

插圖：We begin by looking for the total revenue for ACM. This would be a mea- surement whose enterprise segment is the whole company; that is, the enterprise segment's dimension elements are all at the top of the dimension hierarchies. The measurement normally would not be an absolute value; rather it would be a comparative value with some plan or prior time period. Furthermore, the fact that it is a problem might be indicated by highlighting it according to a ranged phenomenon. The analyst would then begin by looking for problem observations defined by phenomena.To identify that the problem is with equipment sales income, we need to roll back the causal calculation of total revenue as sales income minus sales cost. Note that the causal calculation indicates a possible path of analysis, whether or not the measurement was determined that way. It may be that this final figure was actually sourced from a database. (Due to dirty data, it may be that it doesn't exactly fit the result of the formula.)The next step is to use dimension combination protocols. Looking along the location dimension shows that the Northeast segment had a noticeably higher deviation. We can now focus on the enterprise segment that points to the Northeast dimension element on the location dimension, and at the top for all other dimensions. Repeating this process two more times would lead us to the enterprise segment with location of Northeast, product of 1100 family, and industry sector of government.There is a certain amount of indirection here. When comparative calculations are involved, the route may not be direct. It may not be that the deviation in total revenue is calculated by subtracting the deviation in sales cost from the deviation in sales income. A more likely scenario is that the separate actual and planned sales revenues are calculated, and then these are used in the causal. With absolute deviation, either route will work, but this is not true for percentage deviation. The presence or absence of protocols will indicate what will and will not be appropriate calculations.We can use alternative routes. Instead of first doing the causal, and then dimension combinations, we could break down on the location dimension, then use a causal, and then other dimension breakdowns. There are many possible paths for analysis, and these need not be the same as those used for calculating the figures.

媒體關(guān)注與評(píng)論

“本書對(duì)于模式文化的發(fā)展具有重大貢獻(xiàn)。Martin Fowler 從來(lái)自不同領(lǐng)域的深?yuàn)W的專業(yè)對(duì)象模型中提煉出一組模式，這些模式能幫助你解決具有挑戰(zhàn)性的跨領(lǐng)域建模難題”　　——Erich Gamma 《設(shè)計(jì)模式》作者之一“Martin Fowler給予我們的是答案本身，而不僅僅是尋求答案的過程。就在這樸實(shí)無(wú)華的語(yǔ)言中，你將發(fā)現(xiàn)下一個(gè)業(yè)務(wù)對(duì)象模型的精華所在”　　——Ward Cunningham設(shè)計(jì)模式和敏捷開發(fā)方法先驅(qū)“在這部令人期待己久的著作中，作者為應(yīng)用領(lǐng)域模式所做的工作完全可以與'四人幫'在《設(shè)計(jì)模式》一書中為通用設(shè)計(jì)模式所做的工作相媲美。對(duì)于正在從事面向?qū)ο髽I(yè)務(wù)建?；驑I(yè)務(wù)流程重組的分析與設(shè)計(jì)人員而言，這絕對(duì)是一本必備之書?！薄　　狣onald G. FiresmithSEI軟件大師和咨詢師

編輯推薦

《分析模式:可復(fù)用的對(duì)象模型(英文版)》：業(yè)務(wù)分析是產(chǎn)品和項(xiàng)目開發(fā)的起點(diǎn)，軟件教父Martin Fowler在書中濃縮了十余年的對(duì)象建模經(jīng)驗(yàn)，提煉出覆蓋所有領(lǐng)域的商業(yè)軟件的分析模式和輔助模式，深入講解了如何敏銳地識(shí)別重復(fù)性問題并將其轉(zhuǎn)化為可用模型?！斗治瞿Ｊ?可復(fù)用的對(duì)象模型(英文版)》猶如一個(gè)裝滿妙計(jì)的錦囊。助你輕松應(yīng)對(duì)軟件中各種猝不及防的需求變化。十多年來(lái)。這部經(jīng)典名作在軟件業(yè)的地位無(wú)可替代。它被Ralph Johnson “我一直夢(mèng)寐以求的東西”。每一次閱讀。它都能啟發(fā)新知。和《設(shè)計(jì)模式》一樣。它像一壇陳年佳釀。歷久彌香，值得珍藏與回味。軟件教父的智慧結(jié)晶模式領(lǐng)域的扛鼎之作對(duì)象建模必讀經(jīng)典

圖書封面

圖書標(biāo)簽Tags

無(wú)

評(píng)論、評(píng)分、閱讀與下載

---

    分析模式 PDF格式下載

---