SpecDB

It would be useful for W3C to maintain a database of all of its Technical Reports ("TR documents"), with a history of each. This could be expose this as a bibliographic service, as BibTeX or BIBO or RDF or the format we use for our spec references; the latter would be useful for ReSpec or our other spec-build systems.

This could include a front end to manage them, and an API to allow queries. The front end might include a script to extract information from a spec automatically, which would simplify populating and keeping the database up-to-date.

This could be very useful in our goal of keeping WGs on track. Down the road, we could even make nice visualization tools.

Useful Information

• spec name
• spec shortname
• subtitle
• abstract
• canonical (stable) URL
• anticipated timeline (projected dates for FPWD, LC, CR, PR, Rec)
• each spec version (including Editor's Draft)
  • publication date
  • dated URL
  • alternate URL (e.g., IETF, WHATWG)
  • Recommendation-track status
  • normativity (normative or informative, e.g. spec vs. note)
  • editors
  • authors
  • acknowledgments
  • Working Group
  • comments email
  • bug report URL
  • issue tracker (may be same as bug report URL)
  • Patent Policy terms
  • document license terms
  • translations
    • URL of each translation
  • alternate format (e.g., PDF, ePUB)
    • URL of each alternate format
  • spec dependencies
  • provenance records (eg, reference to the formal and public steps taken for the publication of that particular version, previous versions)

Useful Outputs

• API for queries on document status (internal or external)
• Document schedule and status calendar
• Bibliography or reference generator
• Spec interdependency visualizer

Useful Inputs

• Data extractor: pulls information from a spec to populate database
• Form field, for manual entry

Database

Tables

Please help with this!

Field	Type	Null	Key	Default	extra
spec_id	int(10) unsigned	NO	PRI	NULL	auto_increment
spec_name
spec_shortname

Note: the content should be more structured, reusing other parts of our database. Eg, editor, author, etc, should be references to person's records, and not simply strings.

Types

Just for reference...

• int(size)
• int(size) unsigned
• tinyint(size)
• tinyint(size) unsigned
• bigint(size)
• bigint(size) unsigned
• double
• double unsigned
• binary(size)
• varbinary(size)
• tinyblob
• char(size)
• varchar(size)