From WikiPapers
Jump to: navigation, search

Copied from (some of these notes have been transferred there)

Brief report about the workshop by Phoebe in the Wikimedia Research Newsletter:

Introduction: (cf. )

Reid ...


Existing efforts:

based on:


precedents? not many

  • in CS: graphbib - stopped updating
  • Acawiki
  • Google Scholar


  • filtering/searching/discovering refs
  • reviewing refs
  • managing refs as part of one's workflow


  • sustainability
  • reuse (via an API)
  • watchlists/syndication
  • importing metadata

Use cases Possible use cases for a wiki literature system

Dario: Annotation system to describe relations between papers (e.g. refuting/confirming) Save effort for literature review part in each paper?

Brian: Build tools to identify gaps in the literature or bridge between different communities structure holes collective memories : new "haven't done it" limitations & future suggestions

Tilman different fields segmented by different languages/vocabularies hard to find recent research eg. duplicate research efforts: on quality papers, growth of Wikipedia, user motivation surveys

Han-Teng: resource registry for search engine solve disambiguation and information exploratory with self-reported and -registered entry of resource cover both published work and work-in-progress

Mark: Extending the scope beyond the scholarly literature, e.g. blog postings

Chulki: hard to know where to start and where to stop literature review Extract/build research trends or the evolution of research topics See trends/timelines in research

Paolo: Generate custom feeds of research/watchlists for changes in the literature

(kicking you guys below the fold ;)

Table at the back of the room.

We generated 4 use cases, and for each , we tried to answer the question: why does Google Scholar not cut it for this use case?

Use Case1: As a practitioner I want to get a quick answer on questions like "how much of wikipedia is created by robots?". - Google Scholar doesn't cut it for that need, cause I still need to read all the papers and summarize them. - It would be nice to have a list of 100 or so common questions that come up in the context of wiki research, and have a collaboratively created summary of our current best understanding of that question. ?

Use Case2: As a researcher with an idea for a project, I want to know if it's been done before. - Why doesn't Google Scholar cut it for that? -- If the project idea is somewhat fuzzy, it can involve a lot of digging around in different papers.

Use Case 3: As a wiki researcher, I want to find relevant literature in related fields, that is not necessarily directly associted with wiki. - Why does Google Scholar not cut it here? -- Didn't have time to discuss it.

Use Case 4: As a wiki researcher, I want to know if an article is reputable. - Why does Google Scholar not cut it here? -- Didn't have time to discuss it.

-Watch list daily update citeulike for data exploration - mendalte for data integration

- citeulike

Google Scholar Mendeley Zotero Endnote



Goals/Use Case

  • A Pedagogical activity eg: faculty assigning students to add to it
    • gamification
    • group identification
  • Graduate student to do a visualization, so what should they read?
  • Discover research that is new to you that is more obscure
  • Discussion of an article (why it is important) ***
  • What does a chunk of a discipline look like, chunks at a time
  • Low barrier to participation
  • Dissertation lit review


  • Editability: OK, it is a wiki
  • Personalized Lists/Groups:
  • Watchlist/rss feeds: working on it WikiPapers:RSS
  • Follow new research: sort List of publications by date, or keyword pages
  • Categories/tags/folksonomies: OK, we have categories and keywords
  • Related papers/who cites who: OK, cited by and references sections on every page
  • A good data model (eg: conference v journal v etc): OK, several publication type (conference paper, journal article, thesis, book, etc)
  • Peer-reviewed or not: OK, there is a field in the infobox
  • Export/Import well (correctly?): working on CSV, bibtex, RDF export
  • easy to use: OK, it is a wiki and use Semantic Forms
  • API:
  • finding/citing datasets: OK List of datasets
  • linking to articles: OK
  • encoding future research/shortcomings and linking to those gaps: ?
  • includes many types of literature (e.g. preprints): OK
  • Free/Libre: OK, Creative Commons
  • compatibility: RDF, RSS, plain text
  • indicate open access and/or freely licensed papers: OK, license field
  • recommendation system (related papers, from colleagues, etc): OK, using keywords, cited by, etc
  • personalized stars: ?
  • no wikitext: Semantic Forms is easy to use

Big questions:

  • scope

Next steps:

  • Turn requirements into a table
  • assess existing systems
  • list of research questions
  • Invite pitches for existing systems