UI's for inputting and presenting the metadata of hypermedia documents

1
UI's for inputting and presenting the metadata of
hypermedia documents

• Kai Kuikkaniemi
• 12.11.2002
• HUT T-121.900

2
Idea of the study

• Discuss the role of metadata in creating context
  specific user interface for hypermedia documents
• Theorethical study
• By using a theorethical model, I try to examplify
  the possibilities and importance of metadata

3
Agenda

• Introduction
• Usage of metadata today
• Model
• Usage of metadata tomorrow
• Conclusions

4
Definitions (1)

• Title UI's for inputting and presenting the
  metadata of hypermedia documents
• Inputting
• Where does the metadata come from?
• Presenting
• How is it presented to user
• Metadata
• Definition or description of data

5
Definitions (2)

• Hypermedia document (short documents)
• Orthodoxical view
• Text document with graphics and sounds webpage
• The view in this study
• Limited interactivity (can be said static
  document)
• Multiuser, communicational
• Data structure, components

Text documents
Weblogs
Chat
Mail
Wikies
Presentation
Forms
Simple archives
Static web pages
Collaboration
6
Metadata

• General division of metadata in documents

Semantic
People
Structural
Type
History
Layout
Status
7
Semantic metadata today

• E.g. word definitions
• (Nokia the city, Nokia the company)
• Known problem
• Hard to input in any current (basic e.g. Word)
  document handling applicatations
• Basic way to present is call-out box above the
  word
• Spelling check and thesarus functionality in word
  are related functionalities

8
Status and Type metadata

• Status
• Finished, checked...
• Type
• Document type (report, messaging window)
• Keywords, Class
• Functionality available (hard)
• Same user interface for input and access
• Fairly un-used

9
People and History metadata

• People
• Who am I
• History
• Who, What (changes), When
• This information is mostly included explicitly
  within the document
• Date of last changes updates automatically

10
Structural and layout metadata

• Structural
• Headings, document hierarchy (in general)
• Layout
• How heading looks like, red text
• Tightly bind together
• What you see is what you get
• In Word there is outline (structure) and document
  style (layout) functions

11
General metadata today

• Concept is widely spreaded
• Advantages are acknowledged
• Fairly unused
• Substitute functionalites provided by programs
• Metadata is hard to use
• Most of the metadata types there exists no
  (proper) UIs

12
Model

• Unified hypermedia document management (reading,
  writing, using) tool
• Based on webapplication(s) and modifiable web
  browser
• User manages documents always throw web browser
• Web browser has modifiable (extendable) UI

13
Model outline

• Context specific UI for each document is created
  based on
• Document type (template), history (other users)
• Browser resources (downloaded components)
• User preferences (how likes to edit text, uses
  shortcuts, etc..)
• Document locates always in web server
• Template centric (includes functionality and UI)

14
Model Sequence diagram
Document handler ask user own agent to compose UI
Purpose of requesting document
Document application
User text agent
Document handler
Web Browser
Agents Discuss
Document
Document interface agent
Document content and UI
External UI Components
document content and UI
15
Model Metadata usage
Document handler ask user own agent to compose UI
Purpose of requesting document
Document management application
Type, Status metadata
User text agent
Document handler
Web Browser
Agents Discuss
People metadata
Document
Document interface agent
History metadata
Semantic metadata
Layout metadata
Document content and UI
External UI Components
document content and UI
16
Advantages of the model

• Unified metadata usage
• Documents accessible everywhere
• always published
• Documents more machine understanable
• Easier collaboration
• Correct interface for different usages
• More info (usage of different applications) for
  agent

17
Model in practice

• Forces driving towards this kind of application
• Variety of application, comblicated
• Publishing and collaboration need
• Resource (web, cpu) development
• Business process formalization
• Multidevice, multienvironment
• Search functionalites, semantic web
• Enabling technologies
• XUL, XForms, Phoenix (frontend)
• Wikies, Weblogs (middleware)
• Application server, DMS, PDM (backend)

18
Metadata tomorrow... if the model would be true

• How metadata elements would be generated (input)?
• How can metadata elements be accessed
  (presentting)?
• More metadata is generated automatically, it is
  easier for user to generate more metadata

19
Semantics

• Semantic data is always context specific
• There should be different context specific
  libraries for documents to assign
• Formal (different science vocabulary)
• Informal (e.g. chat language)
• Simple definition for words should included in to
  document (call-out)
• Links ?

20
Status and type

• Status should be generated automatically and/or
  questions from user (E.g. is this finished?)
• Different parts of documents can be in different
  status (processing)
• Status should be explicitely integrated into
  document outlook
• Type should be assigned in the template or in the
  beginning
• Accessing type is relevant already when document
  is searched

21
People and history

• History and people metadata should be generated
  automatically
• People are metadata
• people metadata is important when searching,
  evaluating and accessing document
• History data helps check errors (major undo) and
  understand document progress (for reader)

22
Structural and layout

• For generating structural and layout metadata
  there should be separated functions and views
  available
• Structure and layout are presented implicitly
  when in the document presentation format

23
Conclusions

• Metadata could be the key to
• make document handling easier
• context specific UIs
• structure variety of content
• There is a lot of metadata with different
  purposes, and different accessing and presenting
  rules
• ? Metadata types should be kept separate
• User who generates document is rarely the one who
  access the document metadata
• ? Automatize metadata generation as much as
  possible