PKML: Language Specification, Transforms, Editors, Clients and Visualizations

Personal Knowledge Mark-Up Language (PKML) is an experimental XML implementation designed to support personal knowledgebase, journaling, note-taking, machine learning and agent-based tutorial systems. It is developed by Tim Smith and is utilized in proto-type form in various Designs on Knowledge applications, interfaces and methods.

[Note: this document is in draft form and is used to bring together information described in the DOK project, design and planning. It does not yet represent a complete description of PKML features or projected uses.]

Scenarios (use cases)

Connect the contents of notes and journal entries to Personal Knowledgebase

Automatically record how and when terms and topics were explored or researched

Author self-quizzing, thought questions and trivia

• Creating and embedding learning materials

  • Turn any text into Questions/Answer, Thought Questions

  • Identify new vocabulary and terms. Embed definitions behind the scenes.

  • Self-forming quizzes, integrated with similar content from

    • within the same note

    • other notes within system

    • other sources

  • Track performance, accuracy, frequency

  • Set priority and importance

  • Schedule review

Deepening of note-taking and journaling process

• Embed the results of interactions with deepening, integration, brainstorming processes, and methods

Preprocessing for augmented interactions in

• Natural Language Processing

• Knowledgebase building

• Detecting, authoring, collecting topics, relations, and knowledge

• Quizzing

• Deeping and integration

A higher level description of some scenarios can be found here -> Scenarios

An example scenario can be found here -> Software Environments For “Note-Making” 

Features of PKML

Designed for use in both heterogenous and structured knowledge sources

Embedding of XML trees or data-islands directly into heterogenous documents such as notes and journal entries.

Embedding of more traditional “records” as meta-data

Tag system handles the creation and embedding of the following directly into the text (hidden or visualized as user directs)

• Learning and review resources

• References and citations

• Explicit connections from any idea to any other idea within the same note, other notes, the journal, the knowledge base or external resources.

• Annotation

• SmartTag references to topics and relations within the Personal Knowledgebase

• Creation and visit date time information

  • When was the information added? When was it reviewed?

• To-Do and planning notes and scheduled actions

• Markers to capture the flow and temporal context of notetaking, brain storming, idea formation and research

  • While reading about M in source S, I looked up Y, which led me to idea C. I was thinking about C also because I had recently written about.

• Instructions that allow the author to

  • to meet machine learning utilities part way

  • train personal assistants and agents

  • to drive visualizations and make the knowledge come alive

• Analysis: qualitative and quantitative; automated and manual

• Highlight interesting combinations of words, phrases or terms

• Highlight anything that is meaningful for any reason, with our without explanation

Implementation within DOK software architecture

The data is transformed in some views to expose interactive interface components directly in the rendered document.

Textual analysis Concordance

An PKML editor

Connected to personal knowledge base

Rendering previews

Atomizing (breaking text into constituent parts and structures)

An interactive knowledge authoring interface

Automated and hand editing, combined

Transforms: A family of XSLT Stylesheet transforms to present multiple views of the same data

A library of textual analysis methods optimized for the processing of PKML resources

A system of self-authored prompts used to pose thought questions

Interfaces: authoring, querying and visualization

Other

• Able to handle formatted and heterogenous text

• Islands of information are user defined, both in size and content

• Supports Any Unicode language

• Creating and embedding learning materials

  • Turn any text into Questions/Answer, Thought Questions

  • Identify new vocabulary and terms. Embed definitions

  • Self-forming quizzes, integrated with similar content from

    • Same note

    • Other notes within system

    • other sources

  • Track performance, accuracy, frequency

  • Set priority and importance

  • Schedule review