1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
|
Kascade, an introduction:
Keyword-based search-engines will never have a firm grasp on the huge, growing amount of constantly changing information on the internet. Of course, as time goes by, these engines will get smarter, and eventually probably smart enough for many purposes, yet these or any other 'mechanical' type of search-engine will never be able to give a good picture of the information on the internet related to any specific topic.
Open Directory
Kascade stands for a novel type of search engine, based on the Open Directory principle. This means that anyone can structure a small part of the information on the internet, corresponding to their personal interests or expertise and that the resulting parts are placed in a large categorical structure that others can browse to search for information. Examples of Open Directory initiatives are Dmoz and Infomarker.
Distributed Nature
In contrast with other Open Directory initiatives, though, the parts that people maintain now reside on their own server. What results is a distributed (!) Open Directory. As with Gnutella, though, there are no fixed central servers. This implies that anyone can start a new structure. Actually any piece can be easily replaced, hopefully leading to competitive improvement. The distributed nature and lack of central control facilities make for a system immune to company or government control, much like Gnutella and the Internet itself.
DII File Format
We felt there was another major improvement possible over contemporary Open Directory initiatives. The structures they use are simple hierarchical structures. A more powerful type of structure is possible. We developed a special file format, called DII, that people can use to describe their parts, and that offers extra structural capabilities.
To be able to work with DII files easily, we have developed a graphical editor that can be used to create and maintain them.
( Should you find the following paragraphs to be too technical, you can skip to the Browser section. )
The DII file format features aspects comparable to 'components' and 'functions' as in programming languages, introducing 'abstraction' and thus structure 'reuse' to the Open Directory world. These components and functions themselves may even be distributed! Not less useful, so-called 'queries' may be defined into parts, that query the same and/or other parts, and create new local hierarchical structures in real-time (as users browse the directory.)
DII Example
An example component might be a structure consisting of several DII files distributed over the Internet, that indexes all the countries of the world. Other structures can use this 'country' component to select a country. For instance, a structure about 'culture' might use this component to select a country to go to the structure that describes the culture of a specific country.
The country component 'knows' only where the structure that describes each country is located. The 'culture' structure doesn't know this, so it can use this component to have a user select a specific country. When the user has selected a country, the 'culture' section of the selected structure is entered.
By defining queries in a component, let's say the country component, the user may be enabled to choose among various ways of selecting a country. For instance, the user may first select a continent. Or the user may request the countries be listed in the order of their national income.
As can be seen, the DII file format offers structural capabilities that enable the user to search for information more precisely, hence more efficiently than with contemporary Open Directory initiatives. The DII Specification formally describes the DII file format. We are working on a more accessible tutorial.
Browser
To browse the resulting distributed directories, we have created a special browser, as small as the average ICQ window that can be used to browse for information systematically. Ultimately it fires up a web browser to view the information that has been found. AdvantaAdvantages of having a separate browser for searching are that no big web browser is needed to search and that the user is always confronted with the same, simple layout.
The browser has a built-in IRC client that people can use to chat with each other anywhere in a directory.
Public tool
Finally, kascade is open and free in every sense. Among other things, this mean that:
- it doesn't cost anything to use it, we have no commercial interests
- the source code is publicly available and you can do with it what you want
How you can help
In our download section, the Kascade Editor, Browser and DII Specification may be downloaded. We are currently working on building a basic DII structure, which is used as default by the Browser. You are invited to help us expand this structure, or to compete with us by creating a better structure :-) ( This shouldn't be too hard at this moment! )
|
