Information Categorization

Every categorical distinction is an artificial distinction. The question is: which distinctions (and divisions) are most useful for the purpose at hand?

One way to classify descriptive information:
(1) Description of a state.
(2) Description of a system.
(3) A combination of both.

The whole point of categorizing things is to reduce the number of bits of information you have to carry around in your head. Instead of remembering all the attributes of every item, you simply remember two things:
(1) The category in which the item belongs
(2) The attribute that defines that category.

One way to categorize chunks of information is by how fast they can be retrieved:
(1) In a few seconds, from your own memory.
(2) In a few minutes, from your home or office reference sources (bookshelf, disks, and filing system).

You should not categorize items by a criterion that can vary over time, such as timeliness. Otherwise, you may find that in a few months, the items are in the wrong places.

Words in a dictionary are organized by where their first letters occur in the alphabet. But they don’t have to be organized in this way. They could be organized in many other ways:

(1) alphabetically, by last letter
(2) reverse-alphabetically, by first letter
(3) by the numerical value of the words, with an A=1, B=2, etc.
(4) by subject
(5) by origin
(6) by degree of abstractness

…and dozens if not hundreds more ways. The reason that words are listed alphabetically by their first letters is because it serves a particular purpose in the best way.

Any object can be fit into many different frameworks, each one providing a different perspective on the object. For example, here are some of the frameworks (groups) into which a given man can fit:

• evolutionary line from Australopithecus onward.
• sentient beings
• mammals
• occupation
• genealogical/ancestral line
• country of birth
• citizenship
• tool-using animals
• bipedal animals
• living things
• animals
• things living in the Milky Way galaxy

The search for the best framework (which attribute to use as the sort key).

Problem: how to classify items depicting gross human ignorance?
– Cenohistory, Chronologic?
– Cenohistory, Categoric?
– Human Behaviors file?
– Ideology file?
– Sociology file?

One way to sort my drawings:
(1) Product of intense emotional state
(2) Not a product of intense emotional state

Let a scholar follow his genius. Don’t force him to learn things he doesn’t yet have the knowledge to appreciate.

Sorting papers should be easy. If it’s not, then there’s something wrong with your information processing system.

Beautiful pictures, by definition, are mood-evocative.

Sometimes, simply organizing information in a new way gives you new and valuable insights.

Direct awareness: instant recall on demand.
Indirect awareness: laying down a thread to find your way back to it after you’ve forgotten it.

Sociological items tend to be extremely difficult to sort, because they tend to have many, instead of just one, important categories:
(1) Historical importance
(2) Personal applicability
(3) Subject (a lot of overlap)

It now seems that both discrete (single idea per substrate) and nondiscrete (multiple ideas per substrate) information storage systems both offer their own unique advantages.

Lists are like photographs: they freeze a moment in time. They capture the composition of some aspect of my information management system, many of which are fleeting and temporary, because two kinds of change are always occuring: structural change (change in the system itself) and content change (change in the information passing through the system.

Use notecards if you have:
– many discrete blocks of information (tasks, phone numbers, theory)
– a high turnover rate
– many ways to organize the blocks of information

Use 8.5×11 pages if you have:
– a single, page-sized block of information (map, calendar)
– a low turnover rate
(9/22/85)

Binders are to PIMS what display cases are to museums: showcases of the few, best items, while most are still in storage (files)

The reason some articles are stored (organized) by month, rather than by subject (recreational activity) is because often they will cover more than one subject–the common element being that they are all “partaken of” at the same time of the year.

Items can be sorted in as many ways as there attributes that they have in common.

There are several problems with storing information by how you think you’re going to use it later:
1. You may want to use it in more than one way.
2. You may change the way you use it.
3. You may not be able to predict exactly how you’re going to use it in the future, but you’ve got a strong feeling that it’s going to be useful someday.

The current “Areas”:
1-Career
2-Creative
3-Ideological
4-Personal
5-Recreational
6-School
7-Utilitarian

Currently, my life-areas are:
1. Career/school
2. Cenohistorical
3. Creative
4. Personal
5. Recreational
6. Scideological
7. Utilitarian

“Scideological issues“ and “sociopolitical issues” overlap. In fact, sociopolitical issues are just a collection of related, overgrown scideological issues.

You use categories to make the information manageable. For optimal human comprehension, you want neither too many categories nor too few.

In any categorization scheme, there are two important numbers: the total number of categories and the average number of of items within each category. The goal is to make these numbers as nearly equal as possible. Therefore, a system with four categories of four members each is better than a system of eight categories with two members each (or alternatively, two categories with eight members each). Although each of these systems has 16 members overall, the 4:4 scheme requires the least amount of memory.

All subjects should be at the “basic” level of categorization. That is, neither too high nor too low on the detail hierarchy.

Here is one way to categorize information:
1. That which a moment’s thought will supply.
2. That which can be easily looked up in one place.
3. That which must be synthesized after consulting many sources.
4. That which cannot be looked up anywhere, but through imagination, logic, and intuition, can be figured out.

You should group similar (i.e. same subject) ideas together for the same reason you group similarly-colored puzzle pieces together: it improves your chances of connecting them.

Yet another way to categorize information:
1. Referent information: tells you how to find other information.
2. Terminal information: the end of the search.

Categories. The secret is categories. The trick is to organize the information along the axis of the most appropriate category.

Essay topic: On the Typicality of Various Attributes. Or, How Strong a Link Between Object and Attribute?

It is impossible to categorize one thing. It is also impossible to categorize two things. (Giving them categories would be equivalent to giving them unique names). It takes a minimum of three things in order to assign categories.

Requirements for categories:
1. They must be mutually exclusive. Otherwise you won’t know which of several overlapping categories to search.
2. They must be descriptive of their contents.
3. They must make sense to the user of the categories.

A new way to categorize “multi-attribute” things can be very useful.