群聚的

And those that dwelt in Lacedaemon, lying low among the hills, Pharis, Sparta, with Messe the haunt of doves; Bryseae, Augeae, Amyclae, and Helos upon the sea; Laas, moreover, and Oetylus; these were led by Menelaus of the loud battle-cry, brother to Agamemnon, and of them there were sixty ships, drawn up apart from the others.

来自群山环抱、沟壑宕跌的拉凯代蒙。法里斯、斯巴达和鸽群飞绕的墨塞的兵勇，来自布鲁塞埃和美丽的奥格埃，来自阿姆克莱和濒海的城堡赫洛斯，来自拉斯和俄伊图洛斯地带的兵勇们，由阿伽门农的兄弟、啸吼战场的墨奈劳斯率领，统辖六十条海船，离着其他军旅群聚。

Most of the e-learning research topics focus on metadata of learning object. These studies extract useful information from metadata, and use them to organize the LOs. However, to understand the metadata is not enough to organize the multimedia-style LO in meaningful manner. The comprehensive multimedia-style LO is especially useful for mobile learners to choose what they really demanded. In order to attack the problem, this project uses multi-agent technique to develop a personalized hyper-seamless learning environment for managing LOs comprehensibly. Following that, the proposed environment coordinates two agents: Content Reorganize Agent, and Zone Agent to provide mobile learners the efficient multimedia-style learning experience. The CRA is responsible for reorganizing the discovered critical multimedia content by analyzing learner』s behavior and the multimedia』s annotation. The ZA is designed for terminal devices to monitor the learning behavior, and it sends the collected behaviors to CRA for evaluating. Moreover, the ZAs can communicate with each other, and they can self-organize as a group for sharing learning resources. Due to the cooperation of the two agents, the learning environment can give the properly multimedia-style LO for learners without bandwidth squandering. Finally, we expect this project can construct a comprehensive multimedia-style LO map for efficient personalized seamless learning.

中文摘要大多数的数位学习研究著重於如何利用学习物件的Metadata来组织管理学习物件，然而只透过Metadata仍旧无法有效管理多媒体形态之学习物件，多媒体学习物件的内容若能被了解，则学习者能正确的选择所需的学习内容，尤其是行动学习，能正确选择行动内容不但能减少频宽的浪费，亦能有效利用行动装置有限的资源，有鉴於此，本研究计画拟利用多重代理人机制发展个人化超无缝学习环境来管理内容可被理解之多媒体形态学习物件，在超无缝学习环境中Content Reorganize Agent 与Zone Agent 这两支代理人程式负责协调作业以提供高品质之行动学习经验，其中CRA所负责的工作为分析学习行为与学习物件之注解找出多媒体关键内容来重组具备学习主题之多媒体学习内容，而ZA为安装於学习装置端的代理人程式，其负责监督学生在学习过程中产生的行为，并将该资讯传回CRA以供分析，此外ZA与ZA之间具备自我群聚组织的能力，同一群组下的ZA可彼此共享所需的学习资源，透过CRA与ZA的合作，个人化超无缝学习环境能在不浪费网路频宽的前提下，提供正确且合适的多媒体学习物件给行动学习者，最后我们预期透过本计画的执行能够建立具备内容理解能力之学习物件模型，并利用此模型提供有效率的个人化无缝学习。

According to the domain organization, the GGDEF-EAL containing proteins (pfam database, version 17.0) were classified into A and B types. Six hundreds and five of the 621 proteins were in A type, which was also found in the ancient bacteria Aquifex aeolicus VF5 suggesting that this is the ancestor of the GGDEF-EAL proteins. Phylogenetic analysis of GGDEF- and EAL- containing proteins in the genome of V. vulnificus YJ016 revealed that GGDEF-EAL proteins were clustered. Comparative analysis of domain organization of the clads in the EAL tree allowed identification of several duplication events. The results indicated that most of the GGDEF- and EAL- containing proteins arised by gene duplication.

根据pfam资料库(Version 17.0)，我们整理同时拥有GGDEF和EAL功能性区域的蛋白质，这些蛋白质可以被区分成两种型态，在总数621个GGDEF-EAL中有高达606个是A型态，而且它在古老的细菌Aquifex aeolicus VF5的基因体中也能找到，因此我们推测它可能是GGDEF-EAL的袓先，藉由分析创伤弧菌中所包含之GGDEF和EAL的演化关系，大部份的GGDEF-EAL会群聚在树型相近的位置，进一步比较在EAL树型上有邻近关系的蛋白质序列，根据他们的长度、与所包含的功能区域，我们找到了一些重制的演化痕迹，这样的结果指出，目前看到大多数的GGDEF 和EAL可能都是经过重制产生。

Plunder melds and run with this jewel!

掠夺melds和运行与此宝石！

My dream is to be a crazy growing tree and extend at the edge between the city and the forest.

此刻，也许正是在通往天国的路上，我体验着这白色的晕旋。