WebSchemas/QASchemaResearch

QA schema research

This page provides background research on QA-oriented schemas, beginning with Stack Overflow's data model, as input into WebSchemas and schema.org work on vocabularies for discussion, QA etc.

Stack Overflow data model

See also API docs.

Comments inline on possible representation within schema.org.

Question

This would be a type, subtype of http://schema.org/CreativeWork . See fields marked todo for possible new properties.

• URL (http://schema.org/url exists)
• Title (http://schema.org/name exists)
• Text (http://schema.org/description and http://schema.org/text exist)
• Author (http://schema.org/author exists)
• Votes (Aggregate score, upvotes - downvotes) (http://schema.org/AggregateRating exists; issue: upvotes vs downvotes?)
  • upvoteCount
  • downvoteCount
• Views - todo range = Number
  • viewCount
• Creation Date (http://schema.org/dateCreated exists)
• Last Modified Date (http://schema.org/dateModified exists)
• Last Editor - todo
• Answer Count - todo range = Number (specific to QA?)
  • answerCount
• Comment Count - todo range = Number - long standing issue http://schema.org/UserInteraction inconsistency problems
  • commentCount
• Accepted Answer (if it has one) todo
• Tags (http://schema.org/keywords exists)

Answer

• URL (http://schema.org/url exists)
• Parent URL (always a Question) - todo
• Text (http://schema.org/description and http://schema.org/text exist)
• Author (http://schema.org/author exists)
• Votes (Aggregate score, upvotes - downvotes) (http://schema.org/AggregateRating exists; issue: upvotes vs downvotes?)
  • upvoteCount
  • downvoteCount
• Creation Date (http://schema.org/dateCreated exists)
• Last Modified Date (http://schema.org/dateModified exists)
• Last Editor - todo
• Comment Count - todo range = Number
  • commentCount

Comment (applies to Question or Answer)

Exists as schema.org/Comment

• URL (http://schema.org/url exists)
• Parent URL (Question or Answer) - todo
• Text (http://schema.org/description and http://schema.org/text exist)
• Author (http://schema.org/author exists)
• Votes (Score) (http://schema.org/AggregateRating exists; issue: upvotes vs downvotes?)
  • upvoteCount
  • downvoteCount
• Creation Date (http://schema.org/dateCreated exists)

Numbering properties

We need a convention for naming properties which refer to the number of something (e.g. number of comments). Schema.org feature several inconsistent conventions for naming the properties. See a list of terms which include "num". Among these, the relevant properties are:

numSomething

• http://schema.org/numTracks (MusicPlaylist)

numberOfSomething

• http://schema.org/numberOfEpisodes (TV & radio types)
• http://schema.org/numberOfPages (Book)
• http://schema.org/numberOfSeasons (Series)
• http://schema.org/numberofEmployees (BusinessAudience - error in camelCase but included in this category for this exercise)

somethingCount

• http://schema.org/wordCount
• http://schema.org/offerCount
• http://schema.org/reviewCount
• http://schema.org/ratingCount
• http://schema.org/interactionCount

SIOC uses the num_something convention:

• num_authors
• num_items
• num_replies
• num_threads
• num_views

Proposal for schema.org: adopt the numSomething convention (e.g. numTracks). The following renaming will have to be made:

• numberOfEpisodes -> numEpisodes
• numberOfPages -> numPages
• numberOfSeasons -> numSeasons
• numberofEmployees -> numEmployees

Nearby issues

Wrong range for comment property

out of scope for v1

- favorites