1、<p><b>　　中文3034字</b></p><p>　　Database history </p><p>　　The data bank technology from the birth to the present, in half century's time, has formed the solid rationale, the

2、 mature commercial product and the widespread application domain, attraction more and more researchers joins. The database birth and the development have brought a huge revolution for the computer information management.

3、 More than 30 for years, domestic and foreign already developed have constructed tens of thousands of databases, it has become the enterprise, the department and eve</p><p>　　The Data management of birth<

4、/p><p>　　The database history may trace 50 years ago, at that time data management was simple. Through massive classified, the comparison and the form plan's machine run several million punched card carries

5、 on the data processing, its movement result prints on the paper or makes the new punched card. But the data management is carries on physical to all these punched card the storage and processing.</p><p>　　H

6、owever, Remington Rand Corporation (Remington Rand Inc. ) one kind was called the Univac I computer to promote one kind of one second to be possible to input several hundred records the tape drives, thus has initiated th

7、e data management revolution in 1951. IBM produced first disk driver - the Model 305 RAMAC in 1956. This driver has 50 disk sheets, each disk sheet diameter is 2 feet, may store up 5MB the data. Uses the floppy disk bigg

8、est advantage is may deposit and withdraw the data stochast</p><p>　　Cullinane Corporation (was also afterward Cullinet software company) in 1973, started to sell Goodrich Corporation's IDMS improvement

9、edition, and became at that time in gradually the world the biggest software company.</p><p>　　The Relational database of the origin </p><p>　　The netted database and the level database well ha

10、d already solved the data centralism and the sharing question, but still had in the data independence and the abstract rank is very greatly defective. The user when carries on the deposit to these two kind of databases,

11、still needed to be clear about the data the memory structure, points out the access path. Then appears the relational database has solved these problems well. </p><p>　　In 1970, IBM Researcher Dr. E.F.Codd

12、“Communication of the ACM” on has published named in the publication “A Relational Model of Data for Large Shared Data Banks” the paper, proposed relational model's concept, has laid the relational model rationale. A

13、lthough before Childs already proposed face the set model in 1968, however this paper was revealed generally is in the database system history has the epoch-making significance milestone. The Codd wish is for a database

14、creation exquisite data mod</p><p>　　Structured query language (SQL) </p><p>　　In 1974, IBM Ray Boyce and Don Chamberlin displayed the Codd relational database's 12 criterion's mathemati

15、cs definitions by the simple key words grammar, the milestone type proposed SQL (Structured Query Language) the language. The SQL language's function including the inquiry, the operation, the definition and the contr

16、ol, is one comprehensive, the general relational database language, simultaneously is also one kind of altitude non - proceduring language, only requests the user to point out t</p><p>　　At present the SQL s

17、tandard has 3 editions. The basic SQL definition is ANSIX3135-89, “Database Language - SQL with Integrity Enhancement” [ANS89], is called SQL-89 generally. SQL-89 has defined the mode definition, the data manipulation an

18、d business processes. SQL - 89 and afterward ANSIX3168-1989, “Database Language - Embedded SQL” constituted the first generation of SQL standard. ANSIX3135 - 1992[ANS92] described one kind of enhancement function SQL, no

19、w is called the SQL - 92 standard. SQL -</p><p>　　Object-oriented database </p><p>　　Technology is very mature, but its limitation is also obvious: It can process so-called “the form data” well

20、, more and more complex type's data which appears to the technical fields are actually helpless. After the 90s, the technical fields has been studying and seeks the new database system. But in anything is in the new

21、database system's development direction question, the industrial field once was quite puzzled. At that time technical unrest influence, in quite period of time, people massive </p><p>　　However, several

22、year's development indicated that the object-oriented relations database system product's market development's situation is not ideal. Perfect theoretically has not brought the market the warm response. It no

23、t successful primary cause lies, this kind of database product's main design concept is the attempt substitutes for the existing database system with the new database system. This to many already and accumulated the

24、much work data customer using database system many years, p</p><p>　　Data management of the transformation</p><p>　　In the 1960s later period presented one kind of new database software: Decisio

25、n support program (DSS), its goal is lets the superintendent use the data message effectively in the decision-making process. Therefore in 1970, first on-line analysis processing tool - Express was born. Other decision s

26、upport system follows, many are develops by company's IT department. </p><p>　　In 1985, the first commerce intelligent system (business intelligence) was Procter & Gamble Corporation develops by the

27、 Metaphor computer system Limited company, was mainly uses for to connect the sales message and the retail sales scanner data. In the same year, the Pilot software company starts to sell the first commercial customer/ser

28、ver to carry out information system - Command Center.</p><p>　　Similarly in this year, the UC Berkeley branch school Ingres project evolves Postgres, its goal develops an object-oriented database. Hereafter

29、one year, Graphael Corporation has developed first commercial object database system - Gbase.</p><p>　　In 1988, IBM Corporation's researcher Barry Devlin and Paul Murphy have invented a new terminology -

30、 information warehouse, afterward, the IT manufacturer starts to construct the experimental data warehouse. in 1991, W.H. “Bill” Inmon has published a book “how does construct the data warehouse” the book, causes the dat

31、a warehouse to start to apply truly.</p><p>　　1991: W.H. “Bill” Inmon has published “the construction data warehouse”.</p><p>　　In the 1990s, along with based on the PC customer/server computati

32、on pattern and the enterprise software package's widespread use, data management's transformation completed basically. The data management no longer merely is the memory and the managerial data, but transforms ea

33、ch data management way which the user needs. Internet sudden appearance as well as the XML language's appearance, opened a stretch of new world for the database system's development.</p><p>　　Databas

34、e in the future trend of development </p><p>　　Along with the information management content's unceasing expansion, presented the bountiful data model (level model, network model, relational model, objec

35、t - oriented model, half structurized model and so on), the new technology also emerges one after another incessantly (data stream, Web data management, data mining and so on). At present repeatedly after a few years, in

36、ternationally some senior database expert will gather together, discusses the database research present situation, the new t</p><p>　　In 1989 Future Directions in DBMS Research - The Laguna Beach Participant

37、s.</p><p>　　In 1990 Database Systems: Achievements and Opportunities.</p><p>　　1991:W.H. Inmon has published “Construction Data warehouse” .</p><p>　　1995 Database .</p><

38、p>　　Database development chronicle </p><p>　　1951:Univac system use magnetic tape and the punched card take the data storage.</p><p>　　1956:IBM Corporation first time has introduced disk dri

39、ver. </p><p>　　1961 in its Model 305 RAMAC: GE (GE) Corporation's Charles Bachman has developed first database management system - IDS.</p><p>　　1969:E.F. Codd has invented the relational da

40、tabase.</p><p>　　1973: Led Cullinane Corporation by John J.Cullinane to develop IDMS - one in view of the IBM main engine based on the network model database.</p><p>　　1976: Honeywell Corporatio

41、n has promoted Multics Relational Data the Store-first commercial relational database product.</p><p>　　1979: Oracle Corporation has introduced the first commercial SQL relational database management system.

42、</p><p>　　1983: IBM has promoted the DB2 database product.</p><p>　　1985: For Procter & Gamble system design first commerce intelligent system production.</p><p>　　1991: W.H. “B

43、ill” Inmon has published “the construction data warehouse”.</p><p><b>　　數(shù)據(jù)庫發(fā)展史</b></p><p>　　數(shù)據(jù)庫技術(shù)從誕生到現(xiàn)在，在不到半個世紀的時間里，形成了堅實的理論基礎、成熟的商業(yè)產(chǎn)品和廣泛的應用領域，吸引越來越多的研究者加入。數(shù)據(jù)庫的誕生和發(fā)展給計算機信息管理帶來了一場巨大的革

44、命。三十多年來，國內(nèi)外已經(jīng)開發(fā)建設了成千上萬個數(shù)據(jù)庫，它已成為企業(yè)、部門乃至個人日常工作、生產(chǎn)和生活的基礎設施。同時，隨著應用的擴展與深入，數(shù)據(jù)庫的數(shù)量和規(guī)模越來越大，數(shù)據(jù)庫的研究領域也已經(jīng)大大地拓廣和深化了。30年間數(shù)據(jù)庫領域獲得了三次計算機圖靈（C.W.Bachman,E.F.Codd, J.Gray），更加充分地說明了數(shù)據(jù)庫是一個充滿活力和創(chuàng)新精神的領域。就讓我們沿著歷史的軌跡，追溯一下數(shù)據(jù)庫的發(fā)展歷程。</p>&

45、lt;p><b>　　數(shù)據(jù)管理的誕生</b></p><p>　　數(shù)據(jù)庫的歷史可以追溯到五十年前，那時的數(shù)據(jù)管理非常簡單。通過大量的</p><p>　　分類、比較和表格繪制的機器運行數(shù)百萬穿孔卡片來進行數(shù)據(jù)的處理，其運行結(jié)果在紙上打印出來或者制成新的穿孔卡片。而數(shù)據(jù)管理就是對所有這些穿孔卡片進行物理的儲存和處理。</p><p>　　然

46、而，1951年雷明頓蘭德公司（Remington Rand Inc.）的一種叫做Univac I 的計算機推出了一種一秒鐘可以輸入數(shù)百條記錄的磁帶驅(qū)動器，從而引發(fā)了數(shù)據(jù)管理的革命。1956 年IBM生產(chǎn)出第一個磁盤驅(qū)動器—The Model 305 RAMAC。此驅(qū)動器有50個盤片，每個盤片直徑是2 英尺，可以儲存5MB的數(shù)據(jù)。使用磁盤最大的好處是可以隨機地存取數(shù)據(jù)，而穿孔卡片和磁帶只能順序存取數(shù)據(jù)。</p><p&

47、gt;　　1973 年Cullinane公司（也就是后來的Cullinet軟件公司），開始出售Goodrich公司的IDMS 改進版本，并且逐漸成為當時世界上最大的軟件公司。</p><p><b>　　關(guān)系數(shù)據(jù)庫的由來</b></p><p>　　網(wǎng)狀數(shù)據(jù)庫和層次數(shù)據(jù)庫已經(jīng)很好地解決了數(shù)據(jù)的集中和共享問題，但是在數(shù)據(jù)獨立性和抽象級別上仍有很大欠缺。用戶在對這兩種數(shù)據(jù)

48、庫進行存取時，仍然需要明確數(shù)據(jù)的存儲結(jié)構(gòu)，指出存取路徑。而后來出現(xiàn)的關(guān)系數(shù)據(jù)庫較好地解決了這些問題。</p><p>　　1970年，IBM的研究員E.F.Codd博士在刊物《Communication of the ACM》上發(fā)表了一篇名為“A Relational Model of Data for Large Shared Data Banks”的論文，提出了關(guān)系模型的概念，奠定了關(guān)系模型的理論基礎。盡管之

49、前在1968年Childs已經(jīng)提出了面向集合的模型，然而這篇論文被普遍認為是數(shù)據(jù)庫系統(tǒng)歷史上具有劃時代意義的里程碑。Codd的心愿是為數(shù)據(jù)庫建立一個優(yōu)美的數(shù)據(jù)模型。后來Codd又陸續(xù)發(fā)表多篇文章，論述了范式理論和衡量關(guān)系系統(tǒng)的12條標準，用數(shù)學理論奠定了關(guān)系數(shù)據(jù)庫的基礎。關(guān)系模型有嚴格的數(shù)學基礎，抽象級別比較高，而且簡單清晰，便于理解和使用。但是當時也有人認為關(guān)系模型是理想化的數(shù)據(jù)模型，用來實現(xiàn)DBMS是不現(xiàn)實的，尤其擔心關(guān)系數(shù)據(jù)庫的性

50、能難以接受，更有人視其為當時正在進行中的網(wǎng)狀數(shù)據(jù)庫規(guī)范化工作的嚴重威脅。為了促進對問題的理解，1974年ACM牽頭組織了一次研討會，會上開展了一場分別以Codd和Bachman為首的支持和反對關(guān)系數(shù)據(jù)庫兩派之間的辯論。這次著名的辯論推動了關(guān)系數(shù)據(jù)庫的發(fā)展，使其最終成</p><p>　　結(jié)構(gòu)化查詢語言 (SQL)</p><p>　　1974年，IBM的Ray Boyce和Don Cham

51、berlin將Codd關(guān)系數(shù)據(jù)庫的12條準則的數(shù)學定義以簡單的關(guān)鍵字語法表現(xiàn)出來，里程碑式地提出了SQL(Structured Query Language)語言。SQL語言的功能包括查詢、操縱、定義和控制，是一個綜合的、通用的關(guān)系數(shù)據(jù)庫語言，同時又是一種高度非過程化的語言，只要求用戶指出做什么而不需要指出怎么做。SQL集成實現(xiàn)了數(shù)據(jù)庫生命周期中的全部操作。SQL提供了與關(guān)系數(shù)據(jù)庫進行交互的方法，它可以與標準的編程語言一起工作。自產(chǎn)生之

52、日起，SQL語言便成了檢驗關(guān)系數(shù)據(jù)庫的試金石，而SQL語言標準的每一次變更都指導著關(guān)系數(shù)據(jù)庫產(chǎn)品的發(fā)展方向。然而，直到二十世紀七十年代中期，關(guān)系理論才通過SQL在商業(yè)數(shù)據(jù)庫Oracle和DB2中使用。</p><p>　　目前SQL標準有3個版本?；維QL定義是ANSIX3135-89，“Database Language—SQL with Integrity Enhancement”[ANS89]，一般叫做S

53、QL-89。SQL-89 定義了模式定義、數(shù)據(jù)操作和事務處理。S Q L - 8 9 和隨后的ANSIX3168-1989，“Database Language—Embedded SQL”構(gòu)成了第一代SQL標準。ANSIX3135-1992[ANS92]描述了一種增強功能的SQL，現(xiàn)在叫做SQL-92標準。SQL-92 包括模式操作，動態(tài)創(chuàng)建和SQL語句動態(tài)執(zhí)行、網(wǎng)絡環(huán)境支持等增強特性。在完成SQL-92標準后，ANSI和ISO即開始合

54、作開發(fā)SQL3標準。SQL3的主要特點在于抽象數(shù)據(jù)類型的支持，為新一代對象關(guān)系數(shù)據(jù)庫提供了標準。</p><p><b>　　面向?qū)ο髷?shù)據(jù)庫</b></p><p>　　隨著信息技術(shù)和市場的發(fā)展，人們發(fā)現(xiàn)關(guān)系型數(shù)據(jù)庫系統(tǒng)雖然技術(shù)很成熟，但其局限性也是顯而易見的：它能很好地處理所謂的“表格型數(shù)據(jù)”，卻對技術(shù)界出現(xiàn)的越來越多的復雜類型的數(shù)據(jù)無能為力。九十年代以后，技術(shù)界一

55、直在研究和尋求新型數(shù)據(jù)庫系統(tǒng)。但在什么是新型數(shù)據(jù)庫系統(tǒng)的發(fā)展方向的問題上，產(chǎn)業(yè)界一度是相當困惑的。受當時技術(shù)風潮的影響，在相當一段時間內(nèi)，人們把大量的精力花在研究“面向?qū)ο蟮臄?shù)據(jù)庫系統(tǒng)(object oriented database)”或簡稱“OO數(shù)據(jù)庫系統(tǒng)”。值得一提的是，美國Stonebraker教授提出的面向?qū)ο蟮年P(guān)系型數(shù)據(jù)庫理論曾一度受到產(chǎn)業(yè)界的青睞。而Stonebraker本人也在當時被Informix花大價錢聘為技術(shù)總負責

56、人。</p><p>　　然而，數(shù)年的發(fā)展表明，面向?qū)ο蟮年P(guān)系型數(shù)據(jù)庫系統(tǒng)產(chǎn)品的市場發(fā)展的情況并不理想。理論上的完美性并沒有帶來市場的熱烈反應。其不成功的主要原因在于，這種數(shù)據(jù)庫產(chǎn)品的主要設計思想是企圖用新型數(shù)據(jù)庫系統(tǒng)來取代現(xiàn)有的數(shù)據(jù)庫系統(tǒng)。這對許多已經(jīng)運用數(shù)據(jù)庫系統(tǒng)多年并積累了大量工作數(shù)據(jù)的客戶，尤其是大客戶來說，是無法承受新舊數(shù)據(jù)間的轉(zhuǎn)換而帶來的巨大工作量及巨額開支的。另外， 面向?qū)ο蟮年P(guān)系型數(shù)據(jù)庫系統(tǒng)使查詢

57、語言變得極其復雜，從而使得無論是數(shù)據(jù)庫的開發(fā)商家還是應用客戶都視其復雜的應用技術(shù)為畏途。</p><p><b>　　數(shù)據(jù)管理的變革</b></p><p>　　二十世紀六十年代后期出現(xiàn)了一種新型數(shù)據(jù)庫軟件：決定支持系統(tǒng)(DSS)，其目的是讓管理者在決策過程中更有效地利用數(shù)據(jù)信息。于是在1970年， 第一個聯(lián)機分析處理工具—Express誕生了。其他決策支持系統(tǒng)緊隨其

58、后，許多是由公司的IT部門開發(fā)出來的。</p><p>　　1985年，第一個商務智能系統(tǒng)(business intelligence)由Metaphor計算機系統(tǒng)有限公司為Procter & Gamble公司開發(fā)出來，主要是用來連接銷售信息和零售的掃描儀數(shù)據(jù)。同年，Pilot軟件公司開始出售第一個商用客戶/服務器執(zhí)行信息系統(tǒng)—Command Center。</p><p>　　同

59、樣在這年，加州大學伯克利分校Ingres項目演變成Postgres，其目標是開發(fā)出一個面向?qū)ο蟮臄?shù)據(jù)庫。此后一年，Graphael公司開發(fā)了第一個商用的對象數(shù)據(jù)庫系統(tǒng)—Gbase。</p><p>　　1988年，IBM公司的研究者Barry Devlin和Paul Murphy發(fā)明了一個新的術(shù)語—信息倉庫，之后，IT的廠商開始構(gòu)建實驗性的數(shù)據(jù)倉庫。1991年，W.H. Bill Inmon出版了一本“如何構(gòu)建數(shù)

60、據(jù)倉庫”的書，使得數(shù)據(jù)倉庫真正開始應用。1991： W.H. Bill Inmon發(fā)表了“構(gòu)建數(shù)據(jù)倉庫”。</p><p>　　二十世紀九十年代，隨著基于PC的客戶/服務器計算模式和企業(yè)軟件包的廣泛采用，數(shù)據(jù)管理的變革基本完成。數(shù)據(jù)管理不再僅僅是存儲和管理數(shù)據(jù)，而轉(zhuǎn)變成用戶所需要的各種數(shù)據(jù)管理的方式。Internet的異軍突起以及XML語言的出現(xiàn)，給數(shù)據(jù)庫系統(tǒng)的發(fā)展開辟了一片新的天地。</p>&l

61、t;p><b>　　數(shù)據(jù)庫未來發(fā)展趨勢</b></p><p>　　隨著信息管理內(nèi)容的不斷擴展，出現(xiàn)了豐富多樣的數(shù)據(jù)模型（層次模型，網(wǎng)狀模型，關(guān)系模型，面向?qū)ο竽Ｐ?，半結(jié)構(gòu)化模型等），新技術(shù)也層出不窮（數(shù)據(jù)流，Web數(shù)據(jù)管理，數(shù)據(jù)挖掘等）。目前每隔幾年，國際上一些資深的數(shù)據(jù)庫專家就會聚集一堂，探討數(shù)據(jù)庫研究現(xiàn)狀，存在的問題和未來需要關(guān)注的新技術(shù)焦點。過去已有的幾個類似報告包括：<

62、/p><p>　　1989 年Future Directions in DBMS Research-The Laguna Beach Participants 。</p><p>　　1990 年Database Systems : Achievements and Opportunities。 </p><p>　　1991：W.H.Inmon 發(fā)表了《構(gòu)建數(shù)據(jù)倉庫

63、》。</p><p>　　1995 年的Database。</p><p><b>　　數(shù)據(jù)庫發(fā)展大事記</b></p><p>　　1951：Univac系統(tǒng)使用磁帶和穿孔卡片作為數(shù)據(jù)存儲。</p><p>　　1956：IBM公司在其Model 305 RAMAC中第一次引入了磁盤驅(qū)動器。</p>&l

64、t;p>　　1961：通用電氣(GE)公司的Charles Bachman開發(fā)了第一個數(shù)據(jù)庫管理系統(tǒng)—IDS。</p><p>　　1969：E.F. Codd發(fā)明了關(guān)系數(shù)據(jù)庫。</p><p>　　1973：由John J. Cullinane 領導Cullinane 公司開發(fā)了IDMS—一個針對IBM主機的基于網(wǎng)絡模型的數(shù)據(jù)庫。</p><p>　　197

65、6：Honeywell 公司推出了Multics Relational Data Store—第一個商用關(guān)系數(shù)據(jù)庫產(chǎn)品。</p><p>　　1979：Oracle公司引入了第一個商用SQL關(guān)系數(shù)據(jù)庫管理系統(tǒng)。</p><p>　　1983：IBM 推出了DB2數(shù)據(jù)庫產(chǎn)品。</p><p>　　1985：為Procter & Gamble系統(tǒng)設計的第一個商務