Why can't we organize data well?

Problems In Organizing Information

Problems: - Usual information has various attributes, so it cannot be classified on one criteria. - Associated information should be kept/listed together. - Stored Information should be readily accessible without distraction. - Data should be dynamically structured as desired.

Information management process

1. Collect

2. Analyze

3. Classify <- Problem is here!

4. Archive

5. Browse

Essential elements of information management system

• Structuring <- Problem is here!

• Integrity

• Safety

• Accessibility <- Problem is here!

Existing Methods In Information Management

Keyword Searching

Pros:

• Can be used without having to analyze or classify data

Cons:

• Can not find data if there is no matching keyword in the content of the material

  • E.g., multilingual data, source code, image, sound...

• Unwanted result are retrieved together

• Difficult to find if the keywords are not selected properly

Directory Tree

Pros:

• Ideal for classifying data with single attribute

  • E.g., maps, calendars, organizational chart, etc.

Cons:

• Can't handle Bat Data

  • Let's try to decide where to put 'a bat' -> /HaveWings?, or /Mammal?

• Usual data can not be categorized effectively because its various properties

  • E.g., dates, media types, purposes, subjects, versions, etc.

• Impossible to organize data dynamically with desired criteria

• Difficult to immediate access to desired data

Tagging

Pros:

• Immediate access to data from selecting one of the listed tags

• Well managed data can be classified into various criteria

Cons:

• User have to set several tags into each data manually

• Omitting or mislabeling tags can lead to critical defect

• If the meaning between tags are not mutually exclusive, semantic duplication occurs

• Namespace conflicts when tags have the same title

Introducing `SlipStack`

Data management solution using `Multiple-Classification Tree` method

• Excluding the shortcomings of the directory method and the tag method

• Only the merits of two methods are combined

• Streamlined UX reduces users' time and effort

`Multiple-Classification Tree` method:

• Build up several trees for each attributes in semantically exclusively

  • E.g., Who, What, When, Where, Why, How ... etc.

• Assign nodes to the data, corresponding to the attributes of the data

• Selecting nodes in the trees lists up the result of logical operations of items

  • Filtered by selected nodes (AND, OR, NOT)

Demo Video