Debate tools

From Lesswrongwiki
Revision as of 21:15, 8 February 2010 by Vladimir Nesov (talk | contribs) (sections to structure this would be nice; I suspect this "collaboratively written page" will be a failure like all the others - they all should probably be just prohibited as not expected to succeed.)
Jump to: navigation, search

A debate tool is... any sort of tool that can be used to help somehow with debates...

Sorry, that's a terrible attempt at a definition. Someone please write a better one.

This wiki page was inspired by Debate tools: an experience report

Anyway, this wiki page is for collecting a list of the different debate tools that are available, and for discussing how we can use them.

Or maybe discussion should go in the discussion page, or the comments thread of the original article.

So, here's the list:

  • Argunet
    • summary:
      • Argunet enables you to create argument maps of complex debates online or offline, on your own or in a team.
    • first mentioned:
    • pros:
      • collaboratively edit argument maps
    • cons:
      • not entirely straightforward to use, Morendil had trouble figuring out how to move boxes around.
    • examples:
      • (none yet)




  • Flow
    • summary:
      • a specialized form of note taking called "flowing" within the policy/CEDA/NDT debate community.
    • first mentioned:
    • pros:
      • lots of people have used this technique, and it has been proven to work well
    • cons:
      • it requires a very specific format for the debate
    • examples:
      • (none yet)


  • PyMC
    • summary:
      • a DSL in python for (non-recursive) Bayesian models and Bayesian probability computations.
    • first mentioned:
    • pros:
      • it does Bayesian calculations
    • cons:
      • requires literacy in python and bayesian statistics
    • examples:
      • (none yet)




  • a new tool, developed by LW users, based on MediaWiki, PHP, and GraphViz. and maybe XML
    • summary:
      • a tool that we make ourselves, so that it works the way we want it to work
    • first mentioned:
    • pros:
      • we're writing it, so we can make it work how we want
    • cons:
      • we would need to write it from scratch
    • examples:


  • a new tool, developed by LW users, based on something other than MediaWiki, PHP, and GraphViz.
    • summary:
      • (none yet)
    • first mentioned:
      • (nowhere yet)
    • pros:
      • we're writing it, so we can make it work how we want
    • cons:
      • we would need to write it from scratch
    • examples:
      • (none yet)






Features that an ideal debate tool should have:


  • from almost everyone:
    • an easy to use interface


  • from Morendil:
  • a conclusion or a decision, which is to be "tested" by the use of the tool
  • various hypotheses, which are offered in support or in opposition to the conclusion, with degrees of plausibility
  • logical structure, such as "X follows from Y"
  • challenges to logical structure, such as "X may not necessarily follow from Y, if you grant Z"
  • elements of evidence, which make hypotheses more or less probable
  • recursive relations between these elements


  • from PhilGoetz:
    • an XML-based representation of the data


  • from PeerInfinity
    • generates its results from an annotated log of a debate
    • collaboratively editable, possibly using MediaWiki
    • multiple outfut formats: graphs, tables, the raw data


  • from Johnicholas:
    • Compose in ordinary ASCII or UTF-8
    • Compose primarily a running-text argument, indicating the formal structure with annotations
    • Export as a prettified document, still mostly running text (html and LaTeX)
    • Export as a diagram (automatically layed out, perhaps by graphviz)
    • Export as a bayes net (in possibly several bayes net formats)
    • Export as a machine-checkable proof (in possibly several formats)


  • from Eliezer Yudkowsky:
    • prevents online arguments from retracing the same points over and over.
    • not just graphical with boxes, because that makes poor use of screen real estate.
    • not have lots of fancy argument types and patterns, because no one really uses that stuff
    • a karma system, because otherwise there's no way to find the good stuff.


So, now that everything's all neatly arranged in a list, the next step is to decide whether we want to start using any of these tools, or if we want to create our own.