KM log5 Managing codified knowledge & KM in three organizations

PART I

Manageing Codified Knowledge

1. Even knowledge and expertise that can be shared is often quickly made obsolete.
2. Knowledge about something is called declarative knowledge
Knowledge of how something occurs or is performed is called procedural knowledge.
Knowledge why something occurs is called causal knowledge.
3. Knowledge that is inherently inarticulable yet which firms attempt to make explicit may result in the essence of the knowledge being lost, and performance suffering.
4. Articulable knowledge that has been made explicit represents an exploited opportunity, while leaving inarticulable knowledge in its native form respects the power of tacit knowledge. Both indicate appropriate mangement of the balance between tacit and explicit knowledge.
5. Imagination and flexibility are important, knowledge routinization may be inappropriate.
6. KM is 10% technology and 90% people.

Knowledge Management in three organization: an exploratory study
1. the character of the client and the way that the organization interacted with the client set the framework of the knowledge structures in very important ways.
2. communication culture was very dependent on the management policies adopted by each orgnization, and the commercial nature of their interaction with other organizations.
3. Despite the limited awareness of knowledge management theory and retoric, there was a pervasive understanding of the role of knowledge in ther organization, with some quite well-developmed strategies of embedding knowledge in the organization's operation.
4. the nature of knowledge and knowledge process were intimately related to the nature of the organization, its function, culture, structure and position in the market.
5. the specific nature of the organization, its structure and its specific context must be considered when developing models or theoretical frameworks of kownledge management.

PART II
Both of these articles used practical cases to interpret KM. The first one provided a framwork for configuring a firm's organizational and technical resources capabilities to leverage its codified knowledge. Another one used a exact research method to study different firm's kownledge management.

KM in the first article was separated into 4 parts :
1. the context of knowledge management
2. new organizational roles
3. managing knowledge prcessing applications
4. benefits

The important opinions in this article, I think, are that
1. codified knowledge is explicit knowledge
2. appropriate management of the balance between tacit and explicit knowledge
3. respecting the role of knowledge and learning may be the most effective approach to building a solid and enduring competitive foundation for business organizations.

KM in the second article was separated into 7 parts(in terms of Devenport & Prusak's opinion):
1. the organization: environment and functions
2. governing structures
3. the client
4. knowledge stratege
5. staff skills and development
6. the concept of knowledge
7. information service

The important opinions in this article, I think, are that
1. processes and value of KM are different among organizational culture, commercial structure and other features of each firm.
2. theory and practice need bridge the gap.

PART III
TOPIC: Practices of Knowledge Management

This week, we saw 5 examples: TRI, BL, Law firm, Educational institute, Local Council. In those cases we can realize that there are various dimentions to operate KM, and the dimentions which were choosen by each firms must different in terms of their strategies and organizational ecology.

PART IV

Through those cases mentioned in the articles, we can realize that different organization had different requirements of KM. As the same as the second article said: Knowledge structures and cultures differed substantially between organizations, and were heavily influenced by the commercial enviroment.

The firms, TRI & BL, recited in first article used KM process to develop their KM structure. On contrary, those three firms studied in the second article didn't know KM theory very well, but they still had some activities like KM. This reaffirm the inherence of KM that "KM is an old and new subject". Each organization must have some original activities to treat their own particular knowledge, even they didn't know what KM is. We might consider that KM is an old thing which is rewritten and basic on daily procedure of organizations.

The one makes me most impressive is that "theory and practice need bridge gap", theories usually don't relate practices very well. This situation not only occur in KM, but in many fields.

統計 第五章 間段機率分布

第一節 機率理論
1. 事前機率(古典機率)
假設在一樣本空間有N個樣本點，且每一個樣本點出現的機會相等。某事件A出現的機率為 A/N。

2. 經驗機率
實際重覆的做實驗，將出現次數除以實驗次數。
當實驗的次數越大，經驗機率就越接近先天機率(大數法則)

3. 主觀機率
憑自己的知識與經驗加以猜測。

以上三種理論必須遵守：

• 樣本空間中任一事件的機率不小於0
• 互斥事件聯集的機率就是各事件機率之和
• 樣本空間內所有機率總和為1

第二節 聯合、邊緣、條件機率

1. 聯合機率：
兩個或兩個以上事件同時發生的機率，ex：全部的人之中，吃檳榔又患口腔癌的機率 f(x,y)。

2. 邊緣機率：
只考慮一個樣本空間中事件的發生機率。ex:全部的人之中，吃檳榔的機率 f(y)

3. 條件機率：
固定在一個樣本空間內，另一樣本空間中事件所發生的機率。ex: 在吃檳榔的人之中，患口腔癌的機率 f(x│y)。

條件機率 = 聯合機率/邊緣機率

第四節 間斷機率分布

1. 白努力分布

• 實驗只有兩種結果，成功(機率為p)與不成功(機率為1-p)。
• 平均數 u=p, 變異數=p(1-p)

2. 二項式分布

• 為白努力事件，成功的機率是 p，那麼在n次嘗試中，共成功X次的機率就是二項式分布
• 平均數 = np, 變異數=np(1-p)
• = BINOMDIST(成功次數x, 實驗次數n, 成功機率p, FALSE), FALSE為該次數的機率，TRUE為該次數的累積機率。

3. 負二項分布

• 成功的機率是p，那麼在r次成功之前，已有X次失敗的機率。
• 平均數 = (r/p)-r, 變異數 = r(1-p)/(p平方)
• = NEGBINOMDIST(失敗次數x, 成功次數r, 成功機率p)

4. 超幾何分布

• 一個有限的母體大小為N，其中有M個是成功，如果從這個母體以不放回抽樣法，抽取大小為n的樣本，其中含有X的成功機率
• 平均數 = nM/N, 變異數 = nM(N-M)(N-n)/( N三方 - N平方 )
• = HYPGEOMDIST(要成功的次數X, 樣本大小n, 母體成功的個數M, 母體大小N)
• 當超幾何分布的母體大小N為無限大，就可用二項式分布取代

5. 波氏分布

• 發生於單位時間內的成功次數(入)已知，且成功次數與時間的長短成正比，且兩段不重覆時間內所發生的成功機率是獨立的，且在極短時間內超過一次以上的成功，其機率可以不計。
• 平均數 = 入, 變異數= 入
• = POISSON(事件出現的次數X, 成功的次數(入), FALSE)，TRUE為該次數的機率，FALSE為該次數的累積機率。
• 當二項式分布裡的母數n趨近於無限大，且機率p很小，則二項式分布會趨近波氏分布

6. 間斷均勻分布

KM log4 Knowledge Management and the Dynamic Nature of Knowledge

PATR I
1. Konwledge management or knowledge sharing in organizations is based on an understanding of knowledge creation and knowledge transfer.

2. Knowledge requires knowers, so its processes are interteined with human activity and experience.

3. Knowledge is enriched information with insight into its contxt showing how information and knowledge are closely associated and how they used to define each other.

4. Communicating knowledge is primarily a process, but in order to capture and share knowledge conveniently, its representations are often placed into a storage and retrieval system.

5. One reason knowledge is more valuable than data or information is that it is closer to action.

6. In a kownledge management program it is the knowledge artifact, or the thing, that is managed, not knowledge itself.

7. Instead of the constant initiatives to extract knowledge from within the employees to creat new explicit knowledge artifacts, it might be more productive for organizations to invest effort in creating a kownledge culture.

PART II

This article argue that effective knowledge management in many disciplinary contexts must be based on understanding the dynamic nature of knowledge itself.

The author emphasized that knowledge is dynamic, that is, knowledge is always changing with the human experience and learning. Because of dynamic nature, how to manage knowledge is mentioned on this article. Author said that what in a knowledge management prgram is the "Knowledge Artifact".

And then author addressed three problematic aspects of knowledge management:
1. Knowledge originates and resides in the mind
Separating the mind, body and spirit in defining knowledge and recognizing only the intellectual dimention ignores ignores essential aspects of human nature and presents a fractured picture of knowledge.
As far as knowledge management is concerned about the wholeness of human experience.

2. The technological imperative
IT is just a TOOL, not a solution.

3. Knowledge as a social value
This part mentioned about "organizational knowledge".
Knowledge can make profit if knowledge could be distributed within organization. But it can also be a disadvantage to the organization if it is wrong or if it is inhibiting, or if it is not used for the fulfillment of the organisation's mission.

Anyway, this article gave a good opinion that organizations need to manage knowledge both as "object and process".

PART III

Topic: Knowledge Creation

According to this article, Kownledge is the awareness of what one knows through study, reasoning, experience or association, or through various other types of learning. On the other hand, knowledge is a result of a varid set of prcosses. Through those process, knowledge could be created.

PART IV

"Without person involvment in understanding, knowledge has little value". Although this sentence has been overwrited on other articles, it also make great sense. Activities among people create knowledge and distribute knowledge. If there were no people in there, knowledge would become meaningless.

And I very agree the opinion on "Knowledge Artifact". The last article brought up that knowledge cannot be documented but can be passed through social activities. And this article told us precisely that what we documented or what we handle in KM programs were knowledge artifacts and it is dynamic, changing overtime.

統計 第四章 常態分布

第一節 常態分布的特性

1. 常態分布就是以平均數為中心點，往兩旁漸低的左右對稱分布。常態分布下，中心的最高點就是平均數，也就是眾數、和中位數。
在現實中，並沒有連續的曲線存在，頂多只是類似常態分布，但當樣本數很大時，會越接近常態分布。

2. 常態分布曲線公式(圖4.1)，有平均數和變異數(或標準差)就可知道常態分布的形狀。

• 標準差決定y軸，標準差越小，data越集中
• 平均數決定x軸，平均數不同，圖型會左右位移

3. 讀法：p(X=3)=1/6 → 參數3的出現機率是1/6。只有間斷變項才會有這樣的表達方式，若是連續變項，如身高，就不會說170公分出現的機率是多少。這時候就必須使用「機率密度」

4. 機率密度：

• 適用於連續變項。如平均數170公分，標準差5的常態分布中，170(+-5，165~175)的機率密度為0.0798。但機率密度無多大意義，大家比較關心的是170公分以下的機率，或是165~175的機率。
• = NORMDIST(160, 170, 5, FALSE) → 平均數170，標準差5的常態分布下，160的機率密度。

5. 累積分布函數：

• = NORMDIST(170, 170, 5, TRUE) → 平均數170，標準差5的常態分布下，170以下的累積機率為0.5(50%)。
• 累積分布反函數： 90% = NORMINV(0.9, 170, 5)= 176.41

第二節 標準常態分布 (Z分布)

1. 將平均數定為0，變異數訂為1的常態分布。

2. 將X參數利用線性公式4.25轉換為z分數後，使用 = NORMSDIST(z) 會得到該參數的累積機率。也可利用 = NORMSINV(累積機率) 回求該參數。

第三節 峰度與偏態 (用來描述常態分布的形狀)

1. 常態分布的峰度為0

• 若資料峰度大於0，呈現高峽峰
• 若資料峰度小於零，呈現低闊峰
• KURT(range) 就可得到峰度

3. 偏態

• 偏態值>0，表示資料集中在左邊，右偏態
• 偏態值<0，表示資料集中在右邊，左偏態
• = SKEW(range)

統計 第三章 變異量數與分佈形狀

只用集中量數來描述資料是不夠的，若忽略資料點的分布情形，可能會做出錯誤的判斷。

1. 全距

• 最大值減最小值
• = MAX(range) - MIN(range)
• 優點：容易計算
• 缺點：只用大小值，無法精確反應資料的分布情形

2. 四分位距

• Q=(Q3-Q1)
• 和全距一樣，沒有用到所有資料

3. 平均絕對離差

• 各個數值減掉平均數後絕對值再取平均值
• = AVEDEV(1,2,3,4,5)
• 有絕對值不好計算

4. 變異數

• 平均絕對離差的變異狀，有分為母體變異數和樣本變異數
• 可四則運算，也可推估母體，是推論統計的基石
• 容易受到平均數的極端值影響，因有平方
• 母體 = VARP(range)
• 樣本= VAR(range)

5. 標準差

• 為變異數開根號，分為母體標準差和樣本標準差
• 可四則運算，也可推估母體，是推論統計的基石
• 同樣會受到平均數的極端值影響
• 母體 = STDEVP(range)
• 樣本 = STDEVP(range)

6. 變異係數

• 樣本標準差除以樣本平均數就是變異係數
• 計算標準差或變異數時，每個值都要減去平均數，因此會受到平均數的影響。變異係數可避免過於極端的平均數。
• 變異係數是對於樣本而言的數值，不用來推估母數狀況

KM log3 Knowledge Management: Hype, Hope or help?

PART I

1. Kownledge management, it seems, has two part:

‧there is the management of supporting data and information
‧there is the management of individual with specific abilities.

2. Knowledge is different from data and information, only a person can have and exercise knowledge.

3. Knowledge Management is not so much the management of tangible assets such as data or information, but the active management and support of expertise.

4. Knowledge is not something that can usually be written down, knowledgeable individuals must be encourage to pass their expertise to other through personal contact.

5. For the goal of kownledge worker is not so much to manage kownledge but to solve problems.

PART II

The purpose of this discussion is to look at KM carefully and try to understand what it is, or at least what it could be.
Author used five questions to discuss this topic: KM, hype, hope or help?

Q1: what is knowledge
In this part, author tell different from data, information and knowledge.
He thought Knowledge is not something tangible that we can possess, exchange or lose the way that we can with data or information. And only person can excise it.

Q2: Why are people, especially managers, thinking about knowledge management now?
Every afternoon our corporate knowledge walks out the door and I hope to God they will come back tomorrow. This sentence almost expound why they want to do KM.
And then he started to talk about comminities and IT. He mentioned sharing something is an essential thing although it is hard to creat this kind of culture. And in the IT regard, the author thought the failue of DSS and ES was that people wanted to use them to replace something what can be done only by person.

Q3: What are the enabling technologies for KM?
Store and transmit system. That your workers can find and share something which they want.

Q4: What are the prerequisites for KM?

• Knowledge map
• Knowledge worker

Q5: What are the major challenges for KM?

• culture of sharing
• the treatment of tacit knowledge
• intellecture properties between employees and organization

PART III

Topic: Knowledge and Knowing

This article used "observation (how people use it)" to distinguish the diference between data, information and knowledge. And I use the followed instance to realize these three words.

Before you present something about Tacit Knowledge, you must search a lot of aticles. These articles were composed by words. We can regard these words as DATA. Then I collect these data, and sum up the definition of Tacit Knowledge. This definition is kind of INFORMATION to the others. But for me, I can present it without the summary, so it has turned into KNOWLEDGE. And then I must decide which way to present - by PPT, video or other else - to let the others understand the meaning of Tacit Knowledge, it would be my WISDOM.

Through this instance, we can realize that knowledge is formed from knowing process. Book is not a knowledge until you read it, after this knowing process - reading, something can be internalized become knowledge.

PART IV

Author metioned that knowledge cannot be documented but can be passed through social activities. This opinion is different from last three articles which just mentioned UNDOCUMENTED. And I also agree the opinon about DSS and ES, that is, system can't replace the thing only person can do.

Just as teacher said, the author used words carfully to discuss those questions. But in the conclusion, he told us clearly that Knowledge is not a management but a action. Do something to solve problem. It's realy makes me impressive.

After reading the practical experience of 台積電. I remenber that my friends in StarBucks just felt only StarBucks' coffee is good coffee. They were all assimilated by the company culture, even they had been just hired for 6 months. Consequently, I consider that the culture is not such a difficulte part as the author said. It depends on how deos the culture be distribute in the company.

統計 第二章 集中量數

描述資料的兩種數據：集中量數、變異量數

集中量數(資料集中的情形)
1. 平均數

• 算數平均數：mu=資料為母體，Xbar=資料為樣本。=AVERAGE(1,2,3,4,5,) or = AVERAGE(A1:A5)
• 幾何平均數：適用於平均改變率、平均成長率或平均比率。如近年的經濟成長率為1%,2%,3%，就要用這個公式。=GEOMEAN(5,14,40,125,350)
• 調和平均數：也稱倒數平均數，若資料成等差數列，就使用此平均數。=HARMWAN(80,90)

只有算數平均數才會有母體跟樣本之分，其他則無。

2. 眾數
如果所有的數值都只出現一次，那就沒有眾數，眾數也可能有兩個以上。=MODE(1,1,2,3,4)

3. 中位數
先將數字由大至小排序，中間的數即是中位數。=MEDIAN(1,2,3,4)

4. 截尾平均數
過於極端的值會影響到平均數，所以才將數列排序後，以四分位去掉頭尾，將Q1以下Q3以上的數值排除後，再計算其平均數就是截尾平均數。
Q1= QUARTILE({1,2,3,4,5,6},1)
Q3= QUARTILE({1,2,3,4,5,6},3)

5. 溫賽平均數
與截尾平均數的概念相同，只是溫賽平均數是用四分位數來取代極端值。

量尺的特性

• 名義量尺：眾數
• 順序量尺：眾數、中位數
• 等距量尺：眾數、中位數、平均數

集中量數的優缺點

眾數

• 優點：真實存在、多數意見、容易猜中。
• 缺點：未必能代表集中趨勢

中位數

• 優點：不受極端值影響
• 缺點：不適合四則運算

平均數

• 優點：
• 可以進行四則運算，是推論統計的基礎 。
• 使用了資料所有的數值，因此具有代表性。
• 用平均數來猜測所有數值，產生的誤差最小。
• 樣本的平均數是母體的平均數的最佳估計式（estimator）。
• 平均數較不會受到抽樣變動的影響
• 缺點：
• 容易受到極端值的影響

統計 第一章 序論

第一節 統計學的意義

1. 描述統計學

• 注重資料整理、分析、展示與解釋
• 出處是母體而不是樣本

2. 推論統計學

• 透過整理與分析，藉此推論母體的狀況
• EX: 1萬個燈泡的檢測，樣本必須「充分代表」母體，也就是要隨機取樣，差異性不能太大

3. 樣本數要多大?
見仁見智，求準，樣本就要大。

第二節 變項的分類

1. 質的變項(類別變項)

• 數字無分大小等區別，如宗教、性別
• 必是間斷變項

2. 量的變項

• 如人數、身高等，數字具有量的意義
• 間斷變項：如班級人數，他必是整數
• 連續變項：兩數之間可能有第三數的存在，如身高

3. 其他變項分類
實驗研究：自變項、依變項

第三節 四種測量量尺

1. 名義量尺
性別男1女2，月份1~12(除非限定年份，否則無計算意義)，數字只是代號，沒有順序的差別。

2. 順序量尺

• 有著大小意義
• 如成績90>89，但他不構成等距的條件。因89與88也是差1，但這兩者的1無法描述期間的差異。
• 又如李克特氏量表，非常不同意到非常同意由1~5表示，這也只是順序量尺

3. 等距量尺

• 等距量尺不僅有順序意義，還有差距意義。
• 如攝氏11>10，10>9，這兩者差1度是等量的。

4. 比率量尺

• 最高階的量尺
• 除了有順序、等距意義外，還有「自然零點」。
• 如身高200 cm是100cm的兩倍，換算成其他公制，還是差兩倍
• 但寬鬆一點的說法，如果有個共識基礎，等距量尺就是比率量尺，如溫度雖然會隨公制不同而有差異，但如果共識為攝氏0度，這樣也能達成自然零點的條件

嚴格來說，凡是會使用到平均數和標準差的統計，都不可以使用順序量尺(如智力、成績)。

KM log2

PART I

Artical: What is knowledge management?

1. Concepts are best defined from how people use them.
2. IT-Track KM (management of information) :

• computer and information science
• Knowledge = object

People-Track KM (management of information) :

• philosophy, psychology, sociology......
• Knowledge = Processes

3. Because of their different origins, the two tracks use different language in their dialogues and thus tend to confuse each other when they meet.

4. Anyone can buy a new KM software, but very few have the ability to create sustainable creative organisations.

Part II:

The opinion of IT Track and People Track is come from this article.

IT-Track KM means that the organization focus on computer and information science, and knowledge can be handled by information system. And People-Track KM is that the organization concern about education or training of employees. Learning is the most essencial thing.

The author dislike the notion "KM" personally. He uesed "Knowledge Perspective" or "to be Knowledge Focused" to instead KM. And he seems tended to People Track.

After describing these two tracks, the author started to introduce some companies which take effect on KM.

PART III:

This Week's Topic is : Definition of KM

• In the opinion of Karl, KM includes two tracks.
• In the opinion of Prusak, KM is intergraded by intellectural and practice antecedents.

PART IV:

This week, I learn the development background of KM from Prasuk, and two tracks of KM from Karl. Especially the tracks, it tell difference from two tracks to tell us KM is not only IT but People.

Regarding to the presenter, I very agree her opinion in her conclusion. She said that "you must know what you want to get from knowledge management, and then you can get something from it." As I mentioned in course, companies is too relay on IT to run KM succesfully, especially in Esten countries. They don't really understand the goal of KM, and just set up a lot of system. Then waiting something comes effort automatically. They don't realize that the profit of KM is almost hardly being etimated.

In the conclusion of Prasuk's article, he wanted to see KM would be internalized in organizations, but it is pity that we are still on the origin. We don't make KM into slogan like re-engineer, and we also don't make it into nature. We can't predict what KM will be after 10 years, but we still want to see KM can be naturalized.