I have an idea that sticks with me for a couple of years. It is one that I cannot realize on my own but that I am refusing to let just drop. Because I think bringing it to reality would generate something that could be enlightening, fun and of great value for many people in many different aspects.
The project I would like to propose is a free, collaboratively edited Internet encyclopedia, one that does not compete with Wikipedia but rather perfectly complements it.
This encyclopedia’s main properties are:
The idea is based on two initial thoughts concerning learning and representing knowledge.
Now I would like to present some applications and use cases for the project I am proposing.
Like in Wikipedia you can search for subjects. But instead of getting one large article you just get one node, centered on the screen. The node is represented by its best annotation. You have the options to browse its other annotations, to rate them or add a new one.
Around the node, its most important edges, i.e. relations to other subjects, are shown (with the option to show the other relations too). The more important a subject or relation is, the more highlighted it is rendered.
This is the starting point for an exploration: By clicking on an edge, it will be animated in a cool web 2.0-ish way to the center of the screen, thus replacing the node’s annotations by its own ones. Now you can browse the relation’s annotations, rate them or add a new one.
From there you can go to the relation’s other subject and so on. You might find the most interesting, most enlightening or strangest connections within just a couple of clicks.
To help exploring the data, filters can be applied in a way that only nodes and edges will be shown, that meet certain properties. Those properties may also include semantic logic (see below). You could for example only show subjects and edges with a “consists of” relation type.
This way of searching and exploring differs largely from the static article approach Wikipedia is using and is especially suited for mobile and tablet devices.
My Idea
The project I would like to propose is a free, collaboratively edited Internet encyclopedia, one that does not compete with Wikipedia but rather perfectly complements it.
This encyclopedia’s main properties are:
- Knowledge is represented in a graph. Subjects of knowledge are the nodes; relations between subjects are the edges
- Users can annotate nodes and edges, i.e. write text to explain/describe the subject or relation. In this way relations get to be equally important as subjects.
- An annotation is only editable by a single person. There can be multiple annotations per node/edge however, that can be browsed.
- An annotation is limited in length (e.g. 100 words). This promotes concise texts.
- Annotations can be rated. This ensures that good and concise annotations get promoted.
- Nodes and edges get to be weighted by an algorithm based on the number of annotations and view count.
Main Thoughts
The idea is based on two initial thoughts concerning learning and representing knowledge.
Facts are worthless without relations between them
The first thought is about the importance of facts vs. relations between them. In my opinion relations are at least as important as pure facts. Knowing just facts is practically worthless if you cannot relate them to each other.
Usually, you find out about these relations in the course of a long text, for instance a book or a Wikipedia article, but the longer the text is, the more laborious it is to identify the key relations, which are crucial for understanding an issue.
The project I am proposing wants to separate facts and relations in a way that makes it possible to see and understand the importance of all relations immediately.
Usually, you find out about these relations in the course of a long text, for instance a book or a Wikipedia article, but the longer the text is, the more laborious it is to identify the key relations, which are crucial for understanding an issue.
The project I am proposing wants to separate facts and relations in a way that makes it possible to see and understand the importance of all relations immediately.
Conciseness can be of more value than completeness
The second thought is about the value of conciseness when explaining something.
Have you ever wanted to get to know about a subject, read a long article or even a complete book and at the end, you couldn’t really say what it is really about? This happens to me quite a lot. The overwhelming richness of facts makes it hard to separate what is important and what is not. This can leave you knowing many details but if someone would ask you, what it was about, you couldn’t give a good explanation, not even to yourself.
On the other hand, you sometimes read or hear a short explanation of a subject that you find really enlightening. Many of the facts you knew about it before may suddenly make much more sense because you now have a bigger picture to relate them to. Having such a good short explanation about a subject you want to get to know, can make learning the details by reading long articles or books much easier and more effective.
The project I am proposing wants to promote those short and enlightening explanations.
On the other hand, you sometimes read or hear a short explanation of a subject that you find really enlightening. Many of the facts you knew about it before may suddenly make much more sense because you now have a bigger picture to relate them to. Having such a good short explanation about a subject you want to get to know, can make learning the details by reading long articles or books much easier and more effective.
The project I am proposing wants to promote those short and enlightening explanations.
Applications and Use Cases
Now I would like to present some applications and use cases for the project I am proposing.
Search and Explore
Like in Wikipedia you can search for subjects. But instead of getting one large article you just get one node, centered on the screen. The node is represented by its best annotation. You have the options to browse its other annotations, to rate them or add a new one.
Around the node, its most important edges, i.e. relations to other subjects, are shown (with the option to show the other relations too). The more important a subject or relation is, the more highlighted it is rendered.
 |
| Subject “Thomas Edison” with its most important relations |
This is the starting point for an exploration: By clicking on an edge, it will be animated in a cool web 2.0-ish way to the center of the screen, thus replacing the node’s annotations by its own ones. Now you can browse the relation’s annotations, rate them or add a new one.
 |
| Relation between Thomas Edison and General Electric |
From there you can go to the relation’s other subject and so on. You might find the most interesting, most enlightening or strangest connections within just a couple of clicks.
To help exploring the data, filters can be applied in a way that only nodes and edges will be shown, that meet certain properties. Those properties may also include semantic logic (see below). You could for example only show subjects and edges with a “consists of” relation type.
This way of searching and exploring differs largely from the static article approach Wikipedia is using and is especially suited for mobile and tablet devices.
Find Connections
Because knowledge is represented by a graph, graph algorithms can be applied. Like a route planner calculates the best route between two locations or other websites show you via how many steps you know someone, you could calculate the shortest or best (considering the calculated weight of nodes and edges) connection between two subjects. This could lead to many astounding findings.
Semantic Web
For many years, there has been a lot of talk about the semantic web, but it never really took off. One big problem is getting the semantic data and keeping it up to date. The project I am proposing could easily be enhanced to solve this for many use cases.
A special annotation could be added to every node and every edge. This annotation would contain semantic data and could be edited collaboratively by all users. This way a huge semantic web would be created, one that would constantly grow and be kept up to date.
Given that this special technical annotation is embedded in a non-technical environment – the “normal” annotations are meant to be read by humans not by computers – chances are good the users will maintain the semantic data.
A special annotation could be added to every node and every edge. This annotation would contain semantic data and could be edited collaboratively by all users. This way a huge semantic web would be created, one that would constantly grow and be kept up to date.
Given that this special technical annotation is embedded in a non-technical environment – the “normal” annotations are meant to be read by humans not by computers – chances are good the users will maintain the semantic data.
 |
| Using semantic data in a filter to see Thomas Edison’s connections with cities |
Open Issues
What I am proposing still isn’t much more than a basic idea. There are many things unclear and left to decide. Here is a small list of open issues:
- Evaluate the best usability and user experience (for example by building a prototype).
- Decide of what types the relations may be. Should there be just one universal type or should the relations be directed (unidirectional and bidirectional)? And should relations be allowed to connect more than two subjects?
- What is the best maximum length for an annotation?
- What is the best way to ensure the promotion of good, concise annotations?
- What is the best way to prevent unqualified annotations?
- How can images be included?
Complement to Wikipedia
Like noted before I think the project I am proposing is a perfect complement to Wikipedia.
Wikipedia is a great source for every kind of information. There are new articles and existing ones become better and more detailed every day. This way Wikipedia more and more moves away from being a simple encyclopedia towards a complete reference of all knowledge. But this great development of Wikipedia also has its drawbacks. The overwhelming richness of detailed information can make understanding a subject difficult.
This paves the way for a new, more concise encyclopedia like the one I am proposing. For it can give you a basic scaffold of a subject by providing concise information about what is most important. With this scaffold reading and understanding the Wikipedia article should be a lot easier and efficient.
I think both projects can live peacefully next to each other. The project I am proposing could use the first lines of Wikipedia articles to initially fill subject annotations and both websites could link to each other to leverage each other’s strengths.
Wikipedia is a great source for every kind of information. There are new articles and existing ones become better and more detailed every day. This way Wikipedia more and more moves away from being a simple encyclopedia towards a complete reference of all knowledge. But this great development of Wikipedia also has its drawbacks. The overwhelming richness of detailed information can make understanding a subject difficult.
This paves the way for a new, more concise encyclopedia like the one I am proposing. For it can give you a basic scaffold of a subject by providing concise information about what is most important. With this scaffold reading and understanding the Wikipedia article should be a lot easier and efficient.
I think both projects can live peacefully next to each other. The project I am proposing could use the first lines of Wikipedia articles to initially fill subject annotations and both websites could link to each other to leverage each other’s strengths.
Summery
So this is my idea. Maybe it isn't a very good one but it was bothering me for so long now somewhere in the back of my head that I just had to get it out there somewhere. I hope some of you may find it interesting or maybe even worth realizing. I would very much appreciate your feedback.
