User Tools

Site Tools


Annotation for collaboration

Modelling provenance information for different applications and domains can make distributed collaboration easier.

Identifying documents by their contents makes updating such contents comparatively complex. The strategy in which the URL is used to return the contents does not scale well to a distributed setting where its owner is merely an initial creator. The provenance of a resource can be indicated with web page annotations instead, and ipfs-search could pick this up.

By letting users indicate which resource they are extending, a graph of document versions can be created, similar to how big search engines, like Google and Bing, require users to add meaning to their web pages with Annotations can be added to the published resources using the PROV ontology as a prime vocabulary to indicate document revisions.

  • Leverage Linked Data to create separate indexes for hand-picked types and provide custom interfaces for them.
  • Support discovery of previous/next versions of IPFS resources by scanning for RDFa annotations in the PROV ontology, and building indexes of them.
  • Create documentation, blog posts and videos on how to annotate and extend content on IPFS.


  • Annotation analysis (comparable to anchor-text analysis) can possibly be used in scoring functions
  • For what kinds of low-quality objects might an anchor text/annotation still be of high quality?
  • Is it possible to judge the quality of anchor text/annotations independently of the quality of the rest of the object?
  • Is it possible to detect anchor text/annotation intended to be editorial rather than purely descriptive?
  • In case of documents with multiple topics, can annotations/anchor text analysis be used to identify the themes?
  • Perhaps link-analysis quality judgments can be combined with text-based judgments? A text-based analysis could judge the quality of an object by noting that most near references to the same item attribute other values (a sort of “truth by consensus”, and can also be dangerous of course).

en/problems/pdsearch/annotation.txt · Last modified: 2020/03/09 10:59 by Digital Dot