Published on

A hastily prepared presentation for BarCampNYC3 related to a collaboration platform proposal.

Published in: Technology, Education
  • Be the first to comment

  • Be the first to like this

  1. 1. A proposal for a community collaboration platform using user-driven and machine-driven optimizations. Tuesday, March 18, 2008 1 more information available at or or by e- mailing me at This presentation was written in haste the night before delivery, so some ideas are very likely to be unclear or completely missing. The contents of this file are creative commons licensed (2008, william ward) in whatever default license mode that represents.
  2. 2. Project Goals An open source, freely licensed platform Optimized collaboration by identifying shared needs or shared resources bet ween projects with different or similar goals Project interaction via e-mail, web or API Tuesday, March 18, 2008 2 This is intended to be a platform that is contributed to the community, not a for-profit software product. The mission of the platform is to provide an optimized data analysis tool for identifying overlap in shared goals or requirements. The platform will, of course, provide multiple methods of interacting - including e-mail, a traditional web interface, and a published API preferably following some accepted specification or another.
  3. 3. Theory of Operation Ideas submitted to community, tagged Natural Language Processing applies additional tags Contributors seek shared needs based on weighted scoring of user and machine tags. Tuesday, March 18, 2008 3 The basic flow: An idea or community project is submitted to the platform associated with user-specified tags (a folksonomy, in the current lingo.) The body text description is then sent out to a natural language processing engine, which returns an XML tagged list. Placenames, for example, are tagged as such - as are proper nouns and other concepts in context. Finally, the two “layers” of tags are used with independent weighting to provide computable data for determining similarities (statistically presented - we’re still working in a probability space as far as arbitrary computer linguistics goes for this project.) Contributors can then identify similar opportunities to attach to by using the platform to analyze a “playing field” of projects and grouping them based on similar needs. A project may have many needs, for example - or one may only choose to examine projects with similar goals, or simply analyze the overlap in tags for a coarse similarity approach.
  4. 4. Moving Parts So Far Repository: Semantic Wiki Language Processing: ClearForest SWS UI: Semantic Forms (Yaron Koren) Data Visualization: SIMILE Project (MIT) Tuesday, March 18, 2008 4 The list of potential components is presented on this slide. While not exhaustive, I think these off the shelf resources (or in the case of SIMILE, tools available from this project,) represent viable applications to combine as the platform. I would prefer to use as many solutions provided by the academic community as possible, where substantial research was involved in development.
  5. 5. Challenges Inexperience and time Fundamental feasibility, efficacy Gaming the system for project exposure Maintaining interest Tuesday, March 18, 2008 5 Every project has challenges. Most have time constraints involving the requirement to pay rent, to take consistent paying jobs, and to keep one’s special other suitably happy. This is no different. There may be some naive assumptions about the feasibility of the underlying concept, or the efficacy of this approach versus a simpler - plebeian tagging scheme without the joy of machine dreaming. Gaming the system is a consistent problem in any online service where visibility may yield more donations or contributors. Finally, our online communities have a real problem with short attention spans. Maintaining interest is an art form.
  6. 6. Solutions A good vetting at BarCampNYC for feasibility Meta-moderation to combat abusive gaming Integrating existing tools to minimize required programming Tuesday, March 18, 2008 6 Putting up for a thorough discussion at BarCampNYC would provide a good weeding of pie in the sky concepts. Gaming the system could be combat through meta- moderation, a scheme that works well for Slashdot. Finally, using tools developed by specialists would assist in bringing working code to the project, rather than struggling to implement specific theory with generalist knowledge.
  7. 7. Current state domain exists without content a simple email submission/tracking script is provided by Andy Fundinger totally idle due to competing demands :( Tuesday, March 18, 2008 7 At the time of writing, the domain exists, but has no content. A simple email script exists at which accepts an idea into a database and assigns it a serial number (not a GUID, really, just a simple incrementing integer). The project is idle at this time as most of the participants are actually involved in real startup companies.
  8. 8. What I want... advice and criticism pointers and shortcuts ... that’s about it - Thanks! keep up at or Tuesday, March 18, 2008 8 The point of this presentation was to get advice and criticism of the idea, pointers, and shortcuts to better ways to accomplish these goals. The idea of is free for anyone to implement, but we would like to see it do more than languish. More information will be posted at the URLs above, as things develop. Thanks! - William Ward (March 16, 2008 - BarCampNYC3)