DataFed Tags
From Datafedwiki
<Back to WikiDesign
- Using relational databases to cut down on tags.
Contents |
[edit] Two Different Types of Metadata
Image Annotation on the Semantic Web
- Describing properties of object itself. Many applications use a common, predefined and relatively small vocabulary defining such properties. Examples include the Dublin Core and VRA Core vocabularies.
- Typical annotations provide information about:
- title
- creator
- resolution
- image format
- image size
- copyright
- year of publication, etc.
- Typical annotations provide information about:
- Describing subject matter of the object - what is depicted by the image
[edit] Tagging Methedology
[edit] Two ways to organize content:
1. Hierarchical
2. Database with tag attributes that is searchable/sortable in a long and skinny table.
- Tags need to be well-defined DataFedwiki Tag Conventions
- Tags can be grouped in sets and different sets or combination of sets are used for different page types (Types of Wiki Pages)
[edit] Tag Discipline
The tag-search based navigation, the user may face too many choices, ambiguities in tag naming, and lack of well-defined paths. Note:Must have tag discipline to pick and stick to unique tags, not cluttering the system with multiple tags that mean the same thing.
Add here tag cloud
- Intentional Tags - human decide and excute with intent of tagging. The tagging process can be done either through formal registries that require filling out forms etc. to aid finding content. Alternatively, the tagging can be done ad hoc...
- In Flickr tagging is incidental to sharing pictures which may come naturally - not the purpose. How to make data registration or any other Earth Science tagging to be incidental to something else much more important. Example aerosol events - group hooked on events and content is available and shareable pertaining to events - flickr for AQ events events.
- Interested in organics, collecting stuff labeling with "I'm interested in Organics" Amazon phenomena "people who bought this also bought ..." Main activities like sharing pictures and others are end user-driven.
- another idea here?? Use common interests to attract tagging.
- Rigid tags - not much room for negotiation, consistent (disciplined-Udell) tagging
- Library Analogy (author, publisher, publication date, library location)
- Cyber Analogy: Udell metadata browser(article types, author, dates, tags from del.icio.us/infoworld, URLs)
- Earth Science (Who, Where, When- provider, date, location) - DataFed, OGC, CF)
- Firm tag - parameter
- Library Analogy (Dewey decimal classification - keywords part of hierarchical listing)
- Earth Science (naming conventions e.g. CF Conventions, SOAP namespaces)
- Temporary tags like (page status e.g. InProgress and Done-default)
- Loose tags - flickr, del.icio.us, etc.
- Library Analogy (Keywords)
- Earth Science (users define tags e.g. "Okefenokee Fire" )
- Rigid tags - not much room for negotiation, consistent (disciplined-Udell) tagging
- Unintentional Tag - Usage tagging, resulting from usage.
- last modified date
- size
- how many links to it
- how many times accessed
- Who was here also went there (amazon/itunes)
- Unintentional Tag - Usage tagging, resulting from usage.
Temporary tags 3. Free-form text Search
