I'm working on an idea that the networks that we have today (including the internet) are for host-to-host communication i.e. from one computer to another. As the internet becomes more popular, perhaps the negative significance of this fact will become clearer.

My point is that we naturally communicate one person to the other. A person may be using several computers, or a computer may host several people.

My idea (to be transformed into a project) is a set of free and fast, platform independent programs/standards that enable us to do the things we do today with a very different approach; where your location is less important than who you are.

The idea is to take the concept of the "instant messenger" and pile up various rich capabilities on top of it including rich mail services, equivalent of home pages, etc.

Its still an idea, though ... can anyone conribute to it? Ulimately, on the internet, its people communicating with each other, not machines ... contributions


[ This Message was edited by: seunosewa on 2002-03-31 22:02 ]

That's a cool idea .

Hm, that reminds me of somethingg. We were actually going to make a code-sharing system that uses our own standard. But the project died away eventually.

Sounds difficult...may take years. But it is surely a great idea.

I'm not that good at network programming right now, but that shall improve. I'd like to support this project, even though I think it might prove very difficult. [addsig]

Thanks. I believe the project may actually take years, and will involve several stages. But it will certainly be fun - the brainstorming, the coding, the iteration ... (deep sigh.)

At the lowest conceptual level, the internet as it is today is a network of computers which pass messages to each other (IP packets).

At a higher level, the internet standards (particularly TCP/IP) define what is called a port. A program (which may be referred to as a service) listens on that port for connections. Another program tries to connect to the same port on that same host(computer). When a connection is made, messages are passed back and forth between the two programs involved until the connection is somehow broken or terminated.

So, we can say the internet is a network of programs running on computers and communicating with each other. The problem is that, in the world we have today, a user may be running more than one program on a particular host computer. In fact, a very complex user (for example, a company) may be running several processes on several computers distributed over the world. On the other hand, it is quite possible that a single program acts as the portal allowing thousands of users to communicate (for example, a proxy server). The stress we go through trying to simulate the true state of things (i.e. people, rather than computers, comunicate over the internet) is going to increase as we strive to make our online applicaions (e.g. the YPN site) more interactive and sensitive to individuals.
(...to be continued immediately...)

[ This Message was edited by: seunosewa on 2002-03-31 22:45 ]

THE SOLUTION:
We will consider the internet as a network of people - Human beings, companies, governments, etc. (perhaps the formal term "entities" would be more appropriate).

Since the underrlying network in a network of computers, there must be a computer for each person(entity) that stores information about how that person can be contacted.

I propose this system:

Lets assume I have an entity called xprogrammer and the place to get information about tha entity is http://www.programmers.com. Then, I would advertise my entity name as xprogrammer@www.programmers.com::http. Simple. Anybody wishing to communicate with me will effectively have to download a specified file from http://www.programmers.com which will tell him everything about how to contact me.

Suppose we want to integrate yahoo ids into the system. Then, after working out a way of allowing information about a particular person/entity to be discovered from his yahoo id, we simply define a method of specifying the entity name as a yahoo id. For examply, I use the yahoo id osewa77. We could code this information as osewa77@yahoo.com::yahoo, or simply osewa77@yahoo.

The system of programs (and cross-platform libraries) to be built will determine the method of extracting information about the how to learn about an entity from the Person2Person id and use the yahoo system to get information about the person who owns the id.

Similarly, for passport:
seunosewa@hotmail.com -----> seunosewa@hotmail.com::passport

For information stored on an FTP Server:
seunosewa@ftp.servers.com/seunosewa::ftp

I hope you get the idea -- We're not just coding but defining things.
(...to be continued )

Suppose I request to speak wih an entity person@site.com::http in a messaging program using our system? A file at site.com (perhaps entityinfo.txt) is downloaded. The file will tell the messaging program how best to contact the person for instant communication, among other things. The process is like going to a person's web site to check his AIM or Yahoo! ID, and starting up the appropriate messenger client to comunicate with him if he is online. The diference is that, this time, the process is auomatic.

In addition, there is no need to have a central server to store the contact information. I hope what I'm saying makes a little sense (sigh). I'll try to post more details in a separate site.

Thanks for readng this far

This subject is quite interesting.

One thing I think is a waste of time is the total decentralization, where web pages are connected only through hyperlinks that need to be intepreted by humans.

Imagine a system where each server on the network has a list of "pointers" to other servers with similar content. For example, www.cprogramming.com could point to www.cplusplus.com, www.programmersheaven.com and a few others. Each server has more detailed information (keywords, descriptions) of the inromation on that specific server. This way, it wouldn't take too much effort to check most of the information avaiable in a specific topic, while the network still isn't centralized.

For example, servers could handle specific "keywords". A specific server could have lists for words like "Pascal", "C++" and "COBOL". In the Pascal list, all servers it can find with a "Pascal" keyword are stored. Every time a new server is connected, a message is sent to a computer already connected to the "Pascal" network, and all "Pascal" servers will know of the new computer. To find one "Pascal" computer to start with, we'd need some sort of server that "collects" keywords, i.e. it stores as many servers as it can find for all keywords it can find.

At the moment, I have nothing to code, so I might try to make a simple system like this. Since this idea comes from me, it's most likley bad. However, I think it's interesting.

You've added an interesting dimension to the issue on consideration. The idea of querying bits of information that are similar to each other may fit into the idea of Person2Person communication somehow.

The scheme you mentioned may need to be polished a bit further, since it depends on the cooperation of the web sites being linked. The problem with this is that its difficult to get listened to on such a massive scale if you're not the world wide web consortium. At least its better than the centralized system which would require a superpower like m$ ...

Check out the following links:
http://www.progsnet.org/modules.php?name=News&file=article&sid=9
http://www.progsnet.org/modules.php?op=modload&name=Forum&file=viewtopic&topic=18&forum=13&1
for other thoughts on this same issue.

--------------------------
knowledge is for books,
ideas rule the world,
so lets share ideas together.

This doesn't seem to be anything too new, with lots of IM clients, all the different services are integrated, even other IM protocols, i can send e-mails, SMS, AIM messages, MSN messages etc., visit peoples profiles, home-pages and so on, all from one client. all i need is the sign in name of the person i want to communicate with in some way. Also, i can already guess that the author of http://someshellaccount.org/~user100 can be reached by sending mail to user100@someshellaccount.org.
The internet is already centralised by the existence of DNS, which by no means is a bad thing. Also, if one person corresponds to one computer, what happens when my computer goes down? do i need to pay a DNS fee just to use this service? if people are sending peer to peer messages to me without my permission, that means they know my IP address when i might not want them to. As it is, it isnt really hard to communicate with or find someone over the internet. the only annoyance is that there are so many methods and services ans messengers. Adding a new one won't help this.

I kind of understand what you want to do (or at least I think I do), but what I don't understand is why you would want to do it, or how it would be that different from the way things are now.

This is an interesting idea. It sounds like part of the TUNES Project's goal, because of its decentralized nature.

I recommend that you see how close or far TUNES and, also, metatext systems are to what you wants.

People should be the basic unit of your system IMHO. Organizations are higher level abstractions that, IMHO, aren't *directly* involved at the lowest levels of communication. Acutally the hardware is the lowest level, but does your system model this explicitly? [I suggest you consider TUNES and metatext for their concepts about this kind of thing.] Governments oughtn't even exist. ....

Will graph theory help this project at all, i.e., considering the internet as a big graph? I suspect not but it might...

If you wanted to inform new computers of a certain branch, you have numereous databases, and some scripts on all pages. So, if something is added to one computer, it connects to all the other databases, and adds it into the database. Then, it would make a user accont for the new applicant on every database, the user would only have the bare minimual privalleges they need, to prevent security breaches. The thing is, not many people want to be affilirated completely with other pages. As for the rest of the thread: Standard DNS, at the time been, are perfectly fine with me. I don't know if I fully follow your concept, but if I do, this project is going to need large funding, proffesional and highly experienced programmers and powerful systems. It's definetly an idea, and you could use supernodes like in Kazaa to store information about the networks. Though, I may not fully understand this topic, so all the things I have said may be completely off topic.

Quote:
Imagine a system where each server on the network has a list of "pointers" to other servers with similar content. For example, http://www.cprogramming.com could point to http://www.cplusplus.com, http://www.programmersheaven.com and a few others. Each server has more detailed information (keywords, descriptions) of the inromation on that specific server. This way, it wouldn't take too much effort to check most of the information avaiable in a specific topic, while the network still isn't centralized.


I think the best way to do this would be to make an IE plugin that sends statistics on its users' web browsing to a central server and then extracts correlations, as in, people who visit this website are likely to also visit that website, and then it sends the list of corrolated websites to the user for every website he visits. I think there are already programs that do this but i figured out how to use one.

i thnik it's the best way because it'll naturally give you the most intelligent cross-references (using collective human intelligence), way more powerful than keywords, and it doesn't require that website owners enter keywords.

and not to mention that sites that want popularity would abuse the system by entering as many keywords as possible, although that could be handled by giving a weight directly proportional to the ratio between number keywords you're looking for (one?) and number of keywords you're not.

i can see how this could be integrated with the main idea in this thread. instead of relying on a central server to collect this information, it could be added to some kind of decentralized network that's created to be able to organize a large range of things, from file sharing to person finding to web browsing correlations.. mic/speaker telephone calls, instant messaging, and anything else could be added to it like a module.