Formatting a space

A conventional database design approach will be to create a formatted space to suit algorithms that will be used to sort, search and form relationships between the data items. A human database designer will work out what the categories and sub categories of the formatted areas should be, so that data and information can be inserted into places where the algorithms are designed to expect it.

A living database cannot be constructed in this way because the initial assumption would be that no one has any idea at the start as to what kind of information the living database would be dealing with, let alone knowing all the categories through several levels of subdivision.

My own approach to the construction of a living database was to first create an empty list of twenty-six lines – each of the twenty six lines being identified by the letters of the alphabet: A to Z. This then became the template for the formatting of the entire living database.

When any line is clicked upon, it will either point to a particular area of the formatted space described by the line, or, present another different but identically looking list of twenty-six lines identified by the letters A to Z. These line can then be used as subdivisions of the topic described by the line that has been clicked on. This format all's the multi dimensional space described in the garment example to be created (see figures 8 to 11).

Presenting new lists of lines when categories need to be sub divided allows the database to expand at a second level to a total of 26 times 26, equalling 676 lines. Clicking on any of these 676 lines can present yet another different but identical list of twenty-six lines: allowing a total number of lines at this third level of 26 times 26 times 26, equalling 17,576 lines. Extending this to five levels would allow a total of 11,881,376 lines: creating a structure that can format any space into nearly twelve million different possible topic areas.

Of course, it is extremely unlikely that any area of knowledge would need twelve million categories, but, that is not the point. The idea is that this formatting technique allows ample room in which to expand the content of a living database in many different directions according to the most appropriate way to sub divide it up.

For example, one level may use only two or three of its possible twenty-six sub categories. Many sub categories may not extend to the full five levels. However, with such a large possible space in which the content can extend, the content has the freedom to grow to any size and shape of tree like structure that conveniently divides up the total subject matter. Figure 12 illustrates a typical tree like structure expanding into a formatted space from a first level that uses only four of the possible 26 categories available at that starting level.

Figure 12 shows a database consisting of just four categories at level one. Each of these sub divide. Some of these sub divisions divide several times, creating new levels where at each there might be more sub divisions. Category "A" of the first level through the route of sub divisions "A,A,A, A" points to seven sub divisions at level five.

From this figure, it is easy to see how billions of possible sub division arrangements can be grown in this formatted space, which has twenty-six possible sub divisions at every branch at every level.

The advantage of this system is its elegant simplicity. Any tree like structure that might grow as a result of adding and subdividing categories of the content will be constructed of multiples of exactly the same simple template of twenty-six lines. Using an object oriented environment that can use a templates to create components on demand, there will be no need to create these lists until they are actually needed.

It might seem that there would be a problem in addressing or trying to locate any particular area in a space containing up to 12 million of areas. However, the modular nature of the formatting of the total space allows an elegantly simple solution to this problem. Each of the areas can be described by the route taken to reach the area from the first level.

Addressing any of the twelve million separate possible areas in this space requires a maximum of five letters: the route description from the first level. For example an area at level five in this twelve million area space might be described as "CKRAY". This would indicate that category "C" was chosen at the first level, category "K" chosen at the second, category "R" chosen at the third, category "A" chosen at the fourth and category "Y" chosen at the fifth. This address would identify one particular area out the possible total of 12 million.

If this formatting scheme were used to describe areas where people could meet to discuss a particular topic, this simple addressing system would not only define a particular place to meet to discuss one, out of a possible 12 million topics, it would also describe how anyone could reach that meeting place to insert their presence and to find out who else was there.

With such a simple addressing system, it is possible to string several addresses together to describe being at several meeting places simultaneously. It is also a convenient form of addressing to include in the construction of a bot, so that it knows what places to visit on behalf of its owner.

Notice also that anyone can find a particular topic that interests them, simply by making selections at each level: a maximum of five decision points. This is best explained with an example.

Finding a way through millions of spaces

Trying to visualise a people space with 12 million different possible meeting places might seem impossibly complex. How, for instance, would people find the best areas to establish a bot presence, i.e., find the virtual meeting places where they could make just the right kind of contacts? Surprisingly, it is very easy: it involves nothing more complicated than answering a few multi choice questions: one for each level.

In the hypothetical example described above, merchants could establish a presence at meeting places in a people space – where customers would be looking for the exact garments they were supplying – by answering the following five multi choice questions (selecting by clicking on a single line in each of five given lists – with each click bringing up the list for the next question):

1) What items do you supply? (12 items listed - A to L)

2) What is your geographical location? (12 items listed - A to L)

4) What is the type of locality you supply? (12 items listed - A to L)

3) What languages do you speak? (10 items listed - A to J)

4) How many outlets do you have? (9 items listed - A to I)

The answers to these questions (requiring only five clicks) would automatically create a unique address at one of the possible 155,520 places in the people space. By repeating this simple procedure, a rag merchant could establish a bot presence at every meeting place where there are customers looking for someone like them and wanting to buy the products they were supplying.

In an identical way, buyers can establish their bot presence in the most suitable places to meet the right kind of merchants who are supplying the garments they need– simply by answering five similar multi choice questions (again by clicking on lines in lists):

1) Select the type of garments you are interested in (12 items listed - A to L)

2) Select the geographic areas you can travel to? (12 items listed - A to L)

3) Select the type of locality the supplier comes from? (12 items listed - A to L)

4) Select the languages you speak? (10 items listed - A to J)

5) Select the size of company you want to deal with (number of outlets) (9 items listed - A to I)

Again, each meeting place chosen would require no more than five clicks, each click bringing up the next set of lists. This could be repeated for as many meeting places that a buyer wishes to visit to make contact with appropriate merchants. There will be no sense of the complexity of the formatted space; the users probably won't even realise it exists. They will simply be answering the questions and as a result find themselves, through their bot, automatically connected to the people they need to know.

Programming a personal agent

It is a simple matter for a computer program to record the choices made in these multi choice lists These choices, in the form strings of five letters, can be incorporated in the structure of a personal agent (bot). This will provide the intelligence needed for the personal agent to know where it has to go in a space of 12 million possible areas to represent its owner.

This system automatically matches suppliers to customers. As it is a virtual space, where bot representatives can appear simultaneously in several different places at the same time to cover a variety of needs and wants. Merchants can send their bot representatives to all places where customer bots are looking for the kinds of garments they are supplying

Similarly, a buyer's bot can appear in many different areas if their owners needs a number of different garments, speak several languages, be prepared to travel to several different geographic areas and don't care how many sorters the rag merchant employs. This exploits the unique property of the Internet to allow people to be in multiple virtual places at the same time.

Although this is a hypothetical and probably not a very practical example, it illustrates the essence of the general idea. It is not hard to adapt the conceptual framework to all manner of subject areas where it is useful for people to meet others. This will apply particularly where people need to meet others to be able to share information and knowledge.