(no subject)
I'm fascinated with this medium. Six months ago, I was extremely skeptical of blogging and AIM. Even though I contribute far less to the "blogging community" than others who are truly advancing this medium, I still see endless possibilities for its ability to transform our computers from devices of computation to devices of communication.
I'm enrolled in an advanced distributed operating systems class as part of my Masters in computer science. I have to write a paper exploring some aspect of distributed systems. I've chosen to write about a architecture that integrates the functionality of blogs, AIM, and Friendster, but in an open way that can evolve to incorporate new services. Rather than summarize it here, I'm plonking the proposal below:
The subject of this research paper proposal is not an original analysis of papers covered in this class. Rather, I would like to formally explore the issues related to distributed social networks.
Specifically, I recently joined a service called "Friendster" which has been in the news a lot recently due to its sudden explosive popularity. Each registered user supplies some personal information (biographical, as well as interests, favorite books, etc.). Users can link to one another as "friends" which allow users to discover their friends interests, favorite books, etc., as well as follow links to their friends' friends, and out to infinity. It also has a primitive messaging system, so users can communicate. The only problem is that everyone logs into a single server, subject to frequent outages and slow response. As far as I know, this is the first widely used system for connecting people in this particular way. It allows users to exploit the principle of "six-degrees of separation," which states that any person is at most six acquaintances away from everyone else.
Simultaneously, the practice of weblogging is exploding. People are signing onto websites like livejournal.com, where they supply some personal information (biographical, as well as interests, favorite books, etc.). Users write journals and link to one anothers' journals, which allows the easily distracted to spend hours jumping from 'blog to 'blog, meeting new people (and adding them to a personal "friends" list.) The only problem is that everyone logs into a single server, subject to frequent outages and slow response.
I also use AOL instant messenger to chat with friends when I should be doing research. The only problem is that that system is not linked in any way with the two foregoing systems, which means that users have to explicitly direct their communication partners to a personal webpage, Friendster, or a weblog in order to bring them into their "social circle."
There's a trend here. These are all service-centric systems, not user-centric systems. And if they're made user-centric, the user has a great deal more control of what information is promulgated and to whom, and needs only enter and maintain their information in one place. Each service is implemented as a module in a distributed framework, where each user's runtime environment constitutes a node in a social network.
I envision a system where individuals contract with "web hosting companies" not to host their webpages, but rather as runtime environments where this framework can execute on the Internet. A user's environment will be a store for his personal information, links to friends (links to other users' environments), journal entries, and provide RPC interfaces to get at this information (returned in XML, perhaps, so an information seeker can choose how to render it). Other interfaces can provide messaging services (messages are forwarded to the machine where I'm logged in, or to my telephone).
I won't presume to use this paper as a soapbox for promoting my particular solution to this problem (though I admit to having spent a great deal of time imagining an optimal architecture for this system). Rather, I'd like to take the opportunity to find relevant papers. Specifically, there are many issues of distributed systems regarding security, naming, and many other topics. I'd rather write a survey of some of the available technologies, and analyze how they can be integrated into the social framework concept outlined above.
I fully intend to constrain the scope of this paper, after consulting with one of the professors. I realize that what I've outlined above comprises a huge source of future research.
Of peripheral interest is the fact that though the guidelines of this paper require me to use sources of "academic merit," the sociological aspects of online communication have been explored in much greater depth by those actually involved in making the medium successful. That tends to exclude many people in the academic realm who are where they are because of their inability or unwillingness to communicate...
(Goddammit, I really enjoy writing in this thing. Why don't I do it more often?)
I'm enrolled in an advanced distributed operating systems class as part of my Masters in computer science. I have to write a paper exploring some aspect of distributed systems. I've chosen to write about a architecture that integrates the functionality of blogs, AIM, and Friendster, but in an open way that can evolve to incorporate new services. Rather than summarize it here, I'm plonking the proposal below:
The subject of this research paper proposal is not an original analysis of papers covered in this class. Rather, I would like to formally explore the issues related to distributed social networks.
Specifically, I recently joined a service called "Friendster" which has been in the news a lot recently due to its sudden explosive popularity. Each registered user supplies some personal information (biographical, as well as interests, favorite books, etc.). Users can link to one another as "friends" which allow users to discover their friends interests, favorite books, etc., as well as follow links to their friends' friends, and out to infinity. It also has a primitive messaging system, so users can communicate. The only problem is that everyone logs into a single server, subject to frequent outages and slow response. As far as I know, this is the first widely used system for connecting people in this particular way. It allows users to exploit the principle of "six-degrees of separation," which states that any person is at most six acquaintances away from everyone else.
Simultaneously, the practice of weblogging is exploding. People are signing onto websites like livejournal.com, where they supply some personal information (biographical, as well as interests, favorite books, etc.). Users write journals and link to one anothers' journals, which allows the easily distracted to spend hours jumping from 'blog to 'blog, meeting new people (and adding them to a personal "friends" list.) The only problem is that everyone logs into a single server, subject to frequent outages and slow response.
I also use AOL instant messenger to chat with friends when I should be doing research. The only problem is that that system is not linked in any way with the two foregoing systems, which means that users have to explicitly direct their communication partners to a personal webpage, Friendster, or a weblog in order to bring them into their "social circle."
There's a trend here. These are all service-centric systems, not user-centric systems. And if they're made user-centric, the user has a great deal more control of what information is promulgated and to whom, and needs only enter and maintain their information in one place. Each service is implemented as a module in a distributed framework, where each user's runtime environment constitutes a node in a social network.
I envision a system where individuals contract with "web hosting companies" not to host their webpages, but rather as runtime environments where this framework can execute on the Internet. A user's environment will be a store for his personal information, links to friends (links to other users' environments), journal entries, and provide RPC interfaces to get at this information (returned in XML, perhaps, so an information seeker can choose how to render it). Other interfaces can provide messaging services (messages are forwarded to the machine where I'm logged in, or to my telephone).
I won't presume to use this paper as a soapbox for promoting my particular solution to this problem (though I admit to having spent a great deal of time imagining an optimal architecture for this system). Rather, I'd like to take the opportunity to find relevant papers. Specifically, there are many issues of distributed systems regarding security, naming, and many other topics. I'd rather write a survey of some of the available technologies, and analyze how they can be integrated into the social framework concept outlined above.
I fully intend to constrain the scope of this paper, after consulting with one of the professors. I realize that what I've outlined above comprises a huge source of future research.
Of peripheral interest is the fact that though the guidelines of this paper require me to use sources of "academic merit," the sociological aspects of online communication have been explored in much greater depth by those actually involved in making the medium successful. That tends to exclude many people in the academic realm who are where they are because of their inability or unwillingness to communicate...
(Goddammit, I really enjoy writing in this thing. Why don't I do it more often?)