Page preview panel
ON OFF

This is the graph of pages.

All pages ("nodes") in Knowen belong to a directed acyclic graph: more general nodes are to the left (upstream), and more specific to the right (downstream).

Hover over a node to see the node preview; click to select a specific node; mouse scroll to zoom; click and drag to move.

Dashboard

Now you are in the subtree of Perimeter Open Research project. 

New tools for open research and online collaborations

Many tools and platforms for open research already exist; which features are still missing in these tools? How can we imagine innovations that go beyond ‘just another tool’—that are game changers?

Problem of existence of lots of different tools: by using them, we don’t maximize common knowledge. The more tools we create, the more we dilute this knowledge.

Need a cloud/electronic Kiva!

What works with existing systems

  • Conversational platforms (blogs, SO, etc.)
  • Peer review process? With perhaps some changes, e.g., editors mediate more actively
  • But is an open science ethos compatible with blind review?

Key unmet needs of existing tools

  • Who to trust?
    • How do we develop metrics of “quality of character”? This would encompass…
      • History of contributions (quantity work per project; retractions; some classifications of leaders, contributors, followers)
      • Perhaps a test project to assess interested collaborators, esp. graduate students, post-docs
      • How to measure intrinsic motivation, students who seek opportunities actively?
  • How to democratize access to collaboration networks?
    • How to lower barriers to entry?

Existing tools

  • PeerLibrary, Hypothes.is
  • Fermat’s Library
    • Fermat’s library has a Chrome extension called Librarian for commenting on arXiv papers. If it works that would be great, but it is missing users - I tried it on several maths papers on arXiv but there have been no comments. This may be due to some design problems, or simply lack of exposure to researchers.
  • Github
    • Github is a great tool for open research. The only problems are 1. Steep learning curve for those not familiar with Git / UNIX; 2. No maths rendering support - yes one can host .tex files on Github, but there is no way to have a discussion involving mathematics on Github; 3. Not suitable for hosting data of nontrivial size, but this can be mitigated by hosting data on e.g. Dropbox and providing a link on Github.
  • (what’s missing in) Wikipedia
    • No one has control of the contents. One’s contribution can be easily overwritten by another person
    • Deletionist problem: any entry deemed “not notable enough” risks fast deletion. I imagine many cutting edge research topics can get deleted for this.
    • No attribution/credit or performance measure
    • No original research.
  • Stackexchange / mathoverflow
    • Q&A websites, not meant for discussion like Github issues
    • Questions can be put on hold / locked etc. due to all kinds of reasons even if legitimate
  • Technical discussion
  • Hackernews
    • Great forum for high quality discussions, heavily moderated
    • But not for science research
  • Online document editing: overleaf/sharelatex, google docs
    • Overleaf/sharelatex are just for latex and relatively slow to compile
    • Need to be online to access/edit
    • Google docs is not ideal from a privacy point of view
  • Open knowledge map
    • Use of meta data to create knowledge maps for specific keywords
    • Would be even better with a time series (e.g. how did a discourse develop over time; i.e. new keywords show up and become bigger or a keyword enters a new field …)
  • arXiv
    • Overlay comment systems: SciRate, VoxCharta.
    • overlay journals: SciPost, Quantum
  • Open Journals http://www.theoj.org/
  • Online collaboration/publishing
  • New/better interactive ways to present technical information (read only)

Higher level concepts

  • data mists
  • blockchain republic
  • moon shot

A hacker news for science would be useful. For one thing, we can post links to results of this workshop on it for maximum exposure.

If a tool is to be universally useful (for all scientists), it needs to solve universal problems. Do scientists have universal problems that one tool can solve?