Eclipse 是目前最廣受歡迎的軟體工具整合平台，其可擴充的架構，允許第三方撰寫名為 Plug-in 的擴充套件將既有的或是新開發的工具程式整合至此平台。然而隨著 Eclipse 的普及以及可用工具套件的增加，衍生的困擾則是幾乎每一套 Eclipse 平台都會裝載過多的工具。雖然 Eclipse 的架構提供了諸如 Plug-in 的延遲載入（Lazy-Loading）、快速裝載與更新等功能，讓 Eclipse 平台具有相當高度的可延展性（Scalability），但整體效能還是會因為裝載過多的工具程式而下滑。 為了解決上述問題，本研究提出一套名為 Remote Plug-in 的解決方案，允許使用者或是廠商將原生（Native）或非原生的工具發布到遠端站台，並透過本研究提供的開發工具快速產生包裝程式，讓標的工具能以 Web Services 的型式發布到站台上供使用者叫用。另一方面，需要使用這些工具服務的 Eclipse 平台則可以透過本研究提供的輔助 Plug-in 提供之功能，直接叫用這些工具服務，效果則有如使用安裝於相同電腦的 Plug-in 一樣。 此作法除了能降低 Plug-in 的安裝數量，進而改善 Eclipse 平台的整體效能外，也能減少許多不常使用工具的裝卸。此外，Remote Plug-in 架構的其他優點在於讓用戶得以使用在其平台無法安裝執行的工具程式，軟體商亦可經由發布 Remote Plug-in 加強對軟體的記錄控管。 / Since its inception in 2001, Eclipse has quickly become the most popular platform for software development, partly thanks to its extensible architecture, which allows third parties to easily integrate their software tools into the platform in the form of plug-ins. As the quick increase and abundance of free or cheap Eclipse plug-ins available to the public, coming up also is the problem that most users install too many plug-ins than tolerable in a platform. Although the designers of Eclipse have develped mechanisms such as Lazy-Loading [42], and quick installation and update subsystem to make Eclipse scalable, the overall performance is inevitably degraded for the installation of too many plug-ins. In order to overcome this problem, we propose in this thesis a method called remote plug-ins which can enable the distribution and sharing of plug-ins in remote sites. Users or ISVs (independent software vendors) can deploy their tools on remote sites and use the tool we have developed to wrap them as web services accepting requests from Eclipse platforms. On the other hand, users needing services of remote plug-ins could install in their local Eclipse the proxy plug-in which we developed for requesting services of remote plug-ins. With the help of the proxy, remote plug-ins can be invoked as if they were local ones. By moving unnecessary plug-ins outside of local Eclipse installation, not only can our method improve the overall performance of the platform, but it can also diminish the waste of time on tedious installation and uninstallation of plug-ins which are seldom used. Other advantages of our method include the feasibility of invoking remote plug-ins which are either hard or unable to be installed in local Eclipse, as well as more complete control of vendors to client's access to their tools deployed as remote plug-ins.

網際網路的高速且高度成長，使得今日的網路應用及服務日趨豐富化，網路化理財的觀念，也因此而得以實現。但網路理財牽涉到諸多系統的整合問題，使得目前各金融機構在網路所提供的網路理財服務，在各方面資訊提供及應用上都顯得有限且重複性過高，網路理財的真正效益並沒有真正的達到最大化。網路上各個功能系統的相互獨立，要提供一完整性的解決方案在系統整合上已經是一大瓶頸，再者網路服務的使用者在使用服務時皆受限於網路應用系統本身的功能及資源，使用者無法真正的具備網路服務在選取及使用上的自主性。 　　在本研究中將試圖以XML Web Service的技術架構，建置一整合性的應用服務平台，讓平台能快速的和網路上任一網路應用提供者整合，並能整合網際網路上服務提供者所提供的服務，讓系統和系統間、使用者和系統間，都是以鬆散耦合結構來進行互動。使用者在使用網路服務時也能擺脫過去受限於網站應用提供資及應用上的限制，讓使用者在使用網路服務時能根據自己的需求進行個人化的服務選用，讓使用者得到的是即時性、個人化、整合性的資訊。

本研究旨在先探討新北市國民中學校長分散式領導與教師學術樂觀的現況；接著，剖析不同背景變項與學校變項的教師知覺校長分散式領導與教師學術樂觀的差異情形；再來探討校長分散式領導與教師學術樂觀之間的關係；最後，以逐步多元廻歸檢視校長分散式領導對教師學術樂觀的預測力效果。為達到上述研究目的，本研究採用問卷調查法，以新北市42 所國民中學之主任、組長及教師為研究對象，共寄發問卷895 份，回收有效問卷743 份，有效回收率為83.0％。本研究採用SPSS 20.0 for Windows統計套裝軟體進行分析，獲得以下結論，根據研究結論，提出建議供有關單位參考： 一、新北市國民中學教師知覺校長分散式領導、教師學術樂觀為中高程度。 二、不同教育程度、服務年資、學校地區之教師在知覺校長分散式領導與教師學術樂觀上，沒有顯著差異。 三、51（含）歲組以上的教師知覺校長分散式領導高於30 ~ 40 歲組的教師；在知覺教師學術樂觀則高於所有其他年齡組的教師。 四、學校規模95（含）班以上的教師在知覺教師學術樂觀上，高於44(含)班以下的教師。 五、教師兼主任知覺校長分散式領導高於教師兼組長及教師二組；在知覺教師學術樂觀上，也高於教師兼組長及教師二組。 六、校長分散式領導與教師學術樂觀之間呈現高度正向相關。 七、校長分散式領導對於教師學術樂觀有預測力。 最後，依據上述研究結論，提出具體建議，以做為教育行政機關、國民中學校行政及未來研究人員的參考運用。 / The Study of the Relationships between Principal Distributed Leadership and Teacher Academic Optimism in Junior High School in New Taipei City Abstract First, this study aimed to explore the current situation of New Taipei City junior high schools’ principal distributed leadership and teacher academic optimism; then, to analyze the demographic variables on the different perception of principal distributed leadership and teacher academic optimism, and then to discuss the relationship between principal distributed leadership and teacher academic optimism. Finally, stepwise multiple regression was used to predict teacher academic optimism. Based on the research conclusions, I also proposed recommendations for the relevant institutions’ reference. To achieve these purposes, this study used questionnaire survey method. I used the directors, team leaders and teachers from 42 junior high schools in New Taipei City as my study subjects. A total of 895 questionnaires were distributed, with 743 valid questionnaires. The return rate was 83.0%. Survey data were analyzed by descriptive statistics, t-test, one way ANOVA, Pearson product-moment correlation coefficient, and stepwise multiple regression. The conclusions of this research are as follows: 1. In junior high schools in New Taipei City, teachers’ perception of principal distributed leadership and teacher academic optimism is high. 2. There’s no significant difference in education levels, service years, and school locations in teachers’ perception of distributed leadership and teacher academic optimism. 3. The teachers’ perception over 51 (inclusive) years old on principal distributed leadership is higher than the teachers from 31 to 40. On teacher academic optimism, it’s also higher than all the other groups. 4. The teachers’ perception over 95 (inclusive) classes on teacher academic optimism is higher than the teachers under 44(inclusive) classes. 5. Teachers hold the post as directors whose perceptions both on principal distributed leadership and teacher academic optimism are higher than team leaders and class teachers. 6. Principal distributed leadership is high positively correlated to teacher academic optimism. 7. The prediction of the relationships between principal distributed leadership and teacher academic optimism is supported. Finally, based on the above research conclusions, I provide recommendations as references and applications for educational administrations, junior high school administrations and future researchers.

本研究旨在探討臺北縣國民小學分散式領導與教師學術樂觀的現況，剖析不同背景變項與學校變項知覺分散式領導與教師學術樂觀的差異情形，探討分散式領導對教師學術樂觀的關係，並根據研究結論，提出建議供有關單位參考。為達上述研究目的，本研究採用問卷調查法，以臺北縣60所國民小學之主任、組長及教師為研究對象，共寄發問卷632份，回收有效問卷553份，有效回收率為87.5％。本研究採用SPSS 17.0 for Windows和LISREL 8.80統計套裝軟體進行分析，獲得以下結論： 一、臺北縣國民小學教師知覺分散式領導、教師學術樂觀為中高程度 。 二、不同性別、教育程度、學校規模、學校地區之教師在知覺分散式領導與教師學術樂觀上，沒有顯著差異。 三、50（含）歲以上教師知覺分散式領導高於40歲以下教師，在知 覺教師學術樂觀也高於30（含）~40歲教師。 四、服務年資20（含）年以上教師在知覺分散式領導與教師學術樂觀上，高於10年以下教師。 五、教師兼主任知覺分散式領導高於班級導師；在知覺教師學術樂觀上，也高於教師兼組長及班級導師。 六、學校歷史30年以下與學校歷史90年以上的教師，在知覺分散式領導上，高於學校歷史30（含）~60年及學校歷史60（含）~90年教師；在知覺教師學術樂觀上，亦高於學校歷史30（含）~60年教師。 七、本研究建構之模式經過結構方程模式檢定獲得支持，分散式領導對教師學術樂觀具有正向顯著的影響。 最後依據上述研究結論，提出具體建議，以作為教育行政機關、國小學校行政參考運用。 關鍵詞：分散式領導、教師學術樂觀 / This study aimed to explore the current situation of Taipei County elementary schools’ distributed leadership and teacher academic optimism, and to analyze different background variables and school variables’ on the different perception of distributed leadership and teacher academic optimism, and to discuss the relationship between distributed leadership and teacher academic optimism. Based on the research conclusions, I also submitted recommendations for the relevant institutions’ reference. To achieve these purposes, this study used questionnaire survey method. I used the directors, team leaders and teachers from 60 elementary schools in Taipei County as our study objects. A total of 632 questionnaires were distributed, with 553 valid questionnaires. The effective rate was 87.5%. In this study, I used SPSS 17.0 for Windows and LISREL 8.80 statistical software for analysis and obtained the following conclusions: 1.Taipei County elementary school teachers’ perception of distributed leadership and teacher academic optimism is high. 2.There is no significant difference in different gender, education level, school size, school location in teachers’ perception of distributed leadership and teacher academic optimism. 3.Over 50 (inclusive) years old teachers’ perception on distributed leadership is higher than teachers under 40 years old. The teacher academic optimism is also higher than 30 (inclusive) to 40 years old teachers. 4.Teacher’s year of service above 20 (inclusive) years, their perception on distributed leadership and teacher academic optimism is higher than the teachers under 10 years. 5.Teachers also hold the post as directors’ perception on distributed leadership is higher than class teachers. Their perception of teachers academic optimism is also higher than teachers also hold the post as team leaders and class teachers. 6.Teachers at the school history of less than 30 years and school history over 90 years, their perception of distributed leadership is higher than teachers at school years between 30 (including) years to 60 years and school history of 60 (inclusive) years to 90 years. Their perception of teacher academic optimism is also higher than teachers at school history of 30 (inclusive) years to 60 years. 7.The model constructed in this study was supported through structural equation modeling test. Distributed leadership has a significant positive impact on teacher academic optimism. Finally, based on the above research conclusions, I submitted specific recommendations as references and applications for educational administration and elementary school administration. Keywords: distributed leadership, teacher academic optimism

本研究旨在瞭解國民中學校長分散式領導與學校教育成效現況，並探討國民中學校長分散式領導對學校教育成效之影響。本研究採用調查研究法，自編「分散式領導量表」與「學校教育成效量表」，以八個縣市的公立國民中學正式教師為研究對象，並另外發送校長版問卷以求得校長基本資料。藉由專家效度審查與預試回收有效教師問卷173份，進行項目分析、一階驗證性分析、二階驗證性分析以及信度分析，以形成正式問卷。正式問卷發送1000份教師問卷和100份校長版問卷，回收有效教師問卷567份，回收率為56.7%。本研究以描述性統計、單因子變異數分析、多變量變異數分析、皮爾森積差相關以及結構方程模式等統計方式，對回收資料進行分析。 本研究獲致之結論如下： 一、國民中學校長分散式領導實施現況，尚稱良好。 二、校長分散式領導目前最需加強的是學校「情境」文化部分。 三、校長分散式領導目前表現最好的前三項指標是「校長謙遜有禮」、「校長鼓勵團隊合作」、「校長有清楚的學校願景」。 四、校長分散式領導目前表現最差的三項指標是「本校教職員工勇於嘗試學校新事物」、「本校會依同仁的實際表現來評鑑教職員工的績效」、「本校針對表現不佳的同仁會給予改善建議」。 五、國民中學學校教育成效現況方面，教師部分尚稱良好，學生部分尚可。 六、國民中學學校教育成效，教師成效表現最好的是「教師信任」，最差的是「教師學習」。 七、國民中學學校教育成效方面，學生成效表現最好的是「學生教育機會均等」，最差是「學生學習」。 八、教師因「教師擔任職務」不同，知覺校長分散式領導有顯著差異。 九、教師因「教師性別、教師擔任職務」不同，知覺學校教育成效有顯著差異。 十、教師因「學校規模、校史、學校地區」不同，知覺校長分散式領導有顯著差異。 十一、教師因「學校規模、校史、學校地區」不同，知覺學校教育成效有顯著差異。 十二、教師因「校長性別、校長曾經主要任教領域」不同，知覺校長分散式領導有顯著差異。 十三、教師因「校長性別、校長年齡、校長最高教育程度、校長曾經主要任教領域」不同，知覺學校教育成效有顯著差異。 十四、分散式領導與學校教育成效（學生、教師），均有高度正相關。 十五、校長分散式領導可以有效的預測學校教育成效。 根據研究結果分別提出對教育行政機關、校長以及未來研究的建議如下： 一、對教育行政機關的建議 (一) 宜多辦理教育理念與教育革新議題的相關研習給教師進修。 (二) 舉辦分散式領導相關研習，鼓勵學校成員參與。 (三) 檢視校史達60年以上的學校之教師責任與教師信任。 (四) 建立更有效的學校成員績效評鑑機制。 (五) 加強鄉鎮地區的教育成效關注。 (六) 在學生教育機會均等的進步上，可以參考「導師」的建議。 二、對校長的建議 (一) 校長宜增加與學校成員的互動機會。 (二) 校長可以多鼓勵學校成員嘗試學校新事務。 (三) 學校宜建立更專業發展的回饋機制。 (四) 鼓勵教師主動研究與學習。 (五) 提升學生學習動機。 (六) 關心學生自我概念與具備優良公民美德的能力。 (七) 支持學校成員團隊合作。 三、未來研究的建議 (一) 研究對象：包括不同層級的校長、家長、學生對分散式領導的知覺均為 可能的研究對象。 (二) 研究工具：針對不同研究對象發展不同的問卷。 (三) 研究變項：家長背景亦為可行之研究變項。 (四) 研究方法：加入質性研究加以後續探討。 (五) 統計方法：探討更複雜的分散式領導與各種教育內涵的線性關係。 / The study was to understand the current status of the distributed leadership for junior high principals and effectiveness of school education and to probe into the influences of the distributed leadership for junior high principals on the effectiveness of school education. In this study, I adopted the investigation method, self-prepare the “Distributed Leadership Scale” and the “Effectiveness of School Education Scale”, and took the public junior high school formal teachers in 8 counties/cities as my research samples. Then, I additionally distributed questionnaires to principals for basic information. Through the expert review of validity and pretesting on 173 effective questionnaires, I conducted the item analysis, first order confirmatory factor analysis, second order confirmatory factor analysis, and the reliability analysis to construct the formal questionnaires. Then, 1000 formal questionnaires were issued to the teachers and another 100 questionnaires were issued to the principals with 567 effective teacher questionnaires returned, i.e., 56.7% return rate. Also in this study, I analyzed the returned data by using the descriptive statistics, one-way analysis of variance (or one-way ANOVA), multivariate analysis of variance (or MANOVA), Pearson's correlation and simple linear regression and the structural equation models. The study concludes as follows： 1. The current implementation status of the distributed leadership for junior high principals was moderately good and what needs to be improved most is the “cultural environment” part. 2. The top 3 indexes of the distributed leadership for junior high principals were that “the principal is civil”, that “the principal encourages teamwork”, and that “the principal has a clear vision for the school” while the bottom 3 indexes were that “the faculty staff at this school are brave to try new things”, that “the faculty at this school evaluate each other’s performance based on reality”, and that “the school recommend improvement to the underperforming staff.” 3. For the current status of effectiveness of junior school education, it was moderately good for teachers and above average for the students; Teachers performed best at “teacher trust” but worst at “teacher learning” while students performed best at “equal educational opportunities for students” but worse at “student learning.” 4. For teachers, there was a significant difference in the consciousness of the distributed leadership for principals due to different “teachers positions”, “school size, history, and districts”, and “principal genders and main teaching fields in the past” while there was a significant difference in consciousness of the effectiveness of school education due to different “teacher genders and positions” and “principal genders, ages, and the highest levels of education.” 5. There was a high positive correlation between the distributed leadership and the effectiveness of school education (both for students and teachers) and through the distributed leadership for principals, I could effectively predict the effectiveness of school education. Recommendations for the educational administration, principal, and future research are as follows: Recommendations for the educational administration I recommend the educational administration conduct more workshops concerning further educational believes and innovations for teachers, hold the distributed leadership, encourage school members to participate in relevant workshops, review teacher responsibility and trust at the 60-year-old or older schools, build a more effective school member performance evaluation mechanism, enhance its focus on the educational effectiveness in the rural areas, and refer to “recommendations for the mentors” based on the equality of educational opportunities for the students. Recommendations for the principals I recommend the principals increase opportunities for school member interactions, encourage school members to try new things more, build a more professionally developed mechanism for the students, encourage the teachers to research and learn actively, promote students’ learning motivation, care about student self-concepts and virtues of citizens, and support teamwork at school. Recommendations for future research I recommend the future research include different levels of principals, parents, elder siblings, and students, and their consciousness in distributed leadership in research objects, develop various questionnaires for various research objects in research instruments, include parents and elder siblings in feasible research variables, supplement with qualitative research for further probe in research methods, and probe into the linear relationship between the distributed leadership of more complexity and all kinds of educational contents in statistical methods.

本研究旨在探討國民小學分散式領導對學生學習態度之影響，研究採用調查研究法，以臺北縣國民小學三至六年級級任教師為調查對象，預試問卷，共寄出210份預試問卷，回收199份問卷，回收率94.8%，扣除無效問卷5份，有效問卷為194份，，有效問卷比率92.4%，並以項目分析，二階驗證性因素分析、信度分析進行研究工具之研發。正式問卷發出748份問卷，回收665份問卷，回收率88.9%，扣除無效問卷42份，有效問卷為623份，有效問卷比率83.3%。並分別以平均數、標準差之描述性統計、t考驗、單因子變異數分析、皮爾森積差相關及結構方程模式等統計方式進行資料分析。 本研究獲致以下結論： 一、 國民小學整體分散式領導的現況介於符合和有些符合之間。各層面以成員的參與層面教師知覺程度最好，成員的動力層面較低。 二、 國民小學整體學生學習態度的現況介於符合和有些符合之間。各層面以社會科教師知覺程度最好，英語科較低。 三、 教師人口變項年齡、服務年資不同，知覺校長整體分散式領導達顯著差異。年齡越大，知覺校長分散式領導程度越高。服務年資的組間未達顯著差異。 四、 學校環境變項學校規模、學校區域、校長年資不同，教師知覺校長整體分散式領導達顯著差異。學校規模越小、學校區域偏遠、校長年資越小，知覺校長分散式領導程度越高。 五、 教師人口變項年齡、服務年資不同，教師知覺整體學生學習態度達顯著差異。年齡越大、服務年資越長，知覺整體學生學習態度越高。 六、 學校環境變項學校規模不同，教師知覺整體學生學習態度達顯著差異，事後比較，組間未達顯著差異。 七、 分散式領導與學生學習態度具有中度正相關。 八、 分散式領導對學生學習態度具有正向影響性。 關鍵詞：分散式領導、學習態度、國民小學 / This study aims to explore the relationship between the distributed leadership and students’ learning attitude in the elementary schools. The Survey Research Method is adopted, and the research objects are the third to sixth grade elementary school teachers in Taipei. A total of 210 pre-test questionnaires were distributed, with 194 invalid questionnaires. The effective rate was 92.4%. In this study, item analysis, second-order confirmatory factor analysis, and the reliability analysis are applied. 748 formal questionnaires are distributed, with 665 valid, representing 83.5% effective. With the use of SPSS 17.0 for Windows and LISREL 8.80 statistical software for analysis, descriptive statistics, t-test, one-way analysis of variance, and Person product-moment correlation, and Structural Equation Models were employed in this study, and obtain the conclusion as follows: 1. The overall current conditions of distributed leadership in the elementary schools are good. Objects perform the best in the teacher’s perception of the member’s participation aspect while the member’s motivation performs the worst. 2. The overall current conditions of students’ learning attitude in the elementary schools are good. Among them, teachers in the Social Studies perceived the distributed leadership the best while the ones in the English the worse 3. There is significant difference for different teacher characteristics such as ages, years of service to perceived the principals’ distributed leadership. The older the age, the level of the principal distributed leadership perception the higher. There is no significant difference among the four different year-of-service groups. 4. With the different environmental variables in the school scale, school area, and principal’s years of service, the teachers significantly perceived the principal distributed leadership differently. The smaller the school, the level of principal distributed leadership perception the higher. The same results are obtained if the remoter the school, and the smaller the principal’s years of service, which means the level of principal distributed leadership perception is higher. 5. There is significant difference in the teacher perception of students’ learning attitude with different teacher characteristics such as ages, years of service. The older the teachers’ age, the level of students’ learning attitude perception the higher. 6. With the different scales of schools, the difference achieves significance in the way of teachers sensing the students’ learning attitude. Through post-hoc comparison, there is no difference among groups. 7. Distributed leadership has a significant positive correlation with students’ learning attitude. 8. Distributed leadership has a significant positive impact on students’ learning attitude. Keywords: distributed leadership, learning attitude, elementary school

本篇論文主要在探討如何利用近期發展之深度學習技術在於中文句子分散式表示法學習。近期深度學習受到極大的注目，相關技術也隨之蓬勃發展。然而相關的分散式表示方式，大多以英文為主的其他印歐語系作為主要的衡量對象，也據其特性發展。除了印歐語系外，另外漢藏語系及阿爾泰語系等也有眾多使用人口。還有獨立語系的像日語、韓語等語系存在，各自也有其不同的特性。中文本身屬於漢藏語系，本身具有相當不同的特性，像是孤立語、聲調、量詞等。近來也有許多論文使用多語系的資料集作為評量標準，但鮮少去討論各語言間表現的差異。 本論文利用句子情緒分類之實驗，來比較近期所發展之深度學習之技術與傳統詞向量表示法的差異，我們將以TF-IDF為基準比較其他三個PVDM、Siamese-CBOW及Fasttext的表現差異，也深入探討此些模型對於中文句子情緒分類之表現。 / The paper demonstrates how the deep learning methods published in recent years applied in Chinese sentence representation learning. Recently, the deep learning techniques have attracted the great attention. Related areas also grow enormously. However, the most techniques use Indo-European languages mainly as evaluation objective and developed corresponding to their properties. Besides Indo-European languages, there are Sino-Tibetan language and Altaic language, which also spoken widely. There are only some independent languages like Japanese or Korean, which have their own properties. Chinese itself is belonged to Sino-Tibetan language family and has some characters like isolating language, tone, count word...etc.Recently, many publications also use the multilingual dataset to evaluate their performance, but few of them discuss the differences among different languages. This thesis demonstrates that we perform the sentiment analysis on Chinese Weibo dataset to quantize the effectiveness of different deep learning techniques. We compared the traditional TF-IDF model with PVDM, Siamese-CBOW, and FastText, and evaluate the model they created.

在整個創業環境當中，除了新創公司、政府法令、人才環境等元素以外，投資型機構 也是重要的角色，而所謂創投的分類，也將會因投資標的的不同而有所不同，其中以社會 企業的投資機構，一般我們稱之為「影響力投資機構」，其投資績效不完全以財務報酬為主， 而是會量化所謂社會影響力來做衡量，此機構隨著全世界社會企業的創業潮流，重要性日 益提高，特別的營運模式也開始陸續出現研究。 然而台灣最早在十年前就已有社會影響力投資的概念出現，且相繼出現了許多機構在 推動與執行，產業內具備專業知識與技術的前輩與人才數量更是不勝枚舉，看似不管是資 金還是相關資源皆不缺乏的情況下，回歸到實際執行面上的發展卻不甚順利，許多社會影 響力投資機構相繼地遭遇如營運上的問題，以致於不得不做出像是轉型或是退出市場的決 策，使得台灣有一段時間始終缺乏一個好的典範。一直到 2014 年活水社企投資開發公司(B Current Impact Investment)的正式成立，才真正突破僵局，為台灣的社會影響力投資與社會企業生態再度開啟了新的篇章。 本研究藉由質化訪談的研究模式，實際走入台灣社會企業的現場，面對面蒐集產業面 的經營資訊與競爭策略，暸解作對代表的活水社投是如何辦到的，以時間軸的形式將創立前中後不同期間的經營重點整理並分析，最後輔以經典的投資案例進行必較，將現今台灣最新個案轉為研究資料，最後發現關鍵成功因素為「分散式股權結構」的組織設計模式。 在此架構下的影響力投資機構可以充分發揮各股東的背景與專業能力，互相協調發揮 綜效，同時因應不同的環境與個案條件都能有很好的適應性，這是以往集中式股權結構的 公司無法辦到的，同時他們不斷因應趨勢變化，如今也從個別投資人到公司規模投資機構， 加大了投資基金規模與穩健性，也因此使活水社投得以在不易生存的產業環境下生存下來， 並持續成長茁壯。

Peer-to-Peer(P2P)架構讓社群內的使用者收集分散在網路各處之資源，其中最為風行的P2P系統當屬P2P檔案分享系統。P2P檔案分享系統之架構可分為集中式及分散式，而分散式架構又可細分為結構化及非結構化兩種。採用分散式且非結構化之BitTorrent-like架構，因其可擴張性較佳而廣為風行。在經驗中發現，BitTorrent-like架構在非對稱網路之下雖擁有寬裕的下行頻寬，但是其頻寬使用率卻不高，我們分析其成因如下：(1) Fractional Upward Bandwidth (FUB) 、 (2)Blockage of Acknowledgement (BoA) 與 (3)Long Physical Paths 等。而其中的Blockage of ACK問題，現今尚未有研究學者提出完整解決方案。本研究之目的，即要針對此Blockage of ACK的問題，改良網路協定中的傳輸層 (Transport Layer) 協定以提昇P2P檔案分享系統之效能。我們改用UDP作為傳輸層協定，並在應用層加入自動重建遺失之封包、決定基本傳輸單位大小及決定資料傳送速率等機制，以彌補UDP之缺陷。文中並提供了傳輸協定運作時所需參數的估計法，並且也與其它傳輸協定做效能之比較。實驗結果發現，我們設計的傳輸協定確可改善P2P檔案共享系統的運作效能。 / Peer-to-Peer (P2P) architectures let participants gather resources form network and make participants acquire more computation resources than they could offer. One of the most prominent P2P systems is P2P file sharing system. P2P file sharing system could be classified into to 2 categories: centralized and decentralized model. BitTorrent-like (BT-like) model can be classified as decentralized and unstructured model. BT-like model is quit popular nowadays due to its scalability. Unfortunately, BT-like model has several shortcomings on performance over asymmetric networks, because of some problems, such as Fractional Upward Bandwidth (FUB), Blockage of Acknowledgement (BoA) and long physical paths. No complete solution is available yet to solve the BoA problem. We propose a new UDP-based protocol to alleviate this problem. Since UDP protocol is not able to guarantee the data-integrity nor to determine the proper transmission rate by itself. To offer a complete solution, we have to enhance UDP, to add extra capabilities such as packet loss recovery, segment size determination and data rate determination mechanisms. Experiments show that our proposed protocol has good improvement on performance.

隨著科技日新月異的發展，智慧型手機本身通訊與運算能力也隨著軟體和硬體的改善而不斷地增強，其便利性與高機動性的特色使得越來越多人持有智慧型手機，最後成為人們生活中不可或缺的部份。總觀來說，持有與使用率的上升，不知不覺的形成一種共享經濟與無所不在的行動運算網絡。 基於普及性與相對優秀的運算效能，我們設計與實作出Gnafuy，一個基於行動裝置下的分散式運算框架，希望借用世界上所有閒置行動運算裝置的資源來實行無所不在的運算。 我們發展出一套應用程式介面(API)供開發者依照自己的需求來撰寫自己的分散式運算程式，藉由遵循Gnafuy所制定的應用程式介面，開發者可只專注在演算法本身的開發，而不需要在意其演算法如何被分配到手機上以及待處理資料的分配情形。本篇文章還討論了Gnafuy所採用的分散式運算的程式模型，以及我們如何藉由一個手機應用程式將任務部署至自願者的智慧型手機中，我們發展出一套伺服器端的機制來增加訊息傳遞的成功率，以及偵測計算後回傳結果是否正確，排除被惡意程式污染的客戶端結果。