This is going in warp.

I've read through all the mail about our paper on the portreview
list and attempted to flesh out some main points for our
consideration:


(See also http://lab.bootstrap.org/port)


- universal agreement that the description is too general
  - terms vague (see below)
  - lack of clear example
  - lack of clear problem and goal statement
    - What will a faceted system provide, for who, how will they
      use it
    - But remember: we are not making the facet structure, we are
      making the system to help make it
      - Why is that?
    - Why is access important
      - In general
        - Closed world versus open world
        - Knowledge management versus knowledge generation
        - Archive versus annotation
        - Collaboration as convincing
      - For PORT
      - How is that different from other approaches
  - lack of model of intellectual and theoretical framework
    - augmentation not automation
      - idealogical motivators
      - small tool shims in big process
        - task analysis
        - why humans good for some tasks, why computers good at
          others
    - need to describe
      - large human process
        - how does faceting work now
      - how is our process different
      - labor roadblock of cluster discovery
        - why we think this is responsive to automation (as
          opposed to something else)
  - lack of clear description of user task
    - what is coding
      - any reports of practice
    - what is traversal
    - what do you get when it is all done

- definitions and use of terms imprecise
  - especially
    - cluster
    - facet
      - this is the cruz of the biscuit
    - access structure
      (i think we should take care to distinguish between acces
       structures for retrieval (searching, browsing, class
       structures) and for reference (purple etc))
    - coding messages
    - term compared to concept
      - these are different
      - how?
      - use needs to be precise in paper

- Classic indexation system imprecise (Martin's primary bone)
  - Does not lead to real "concept" or "structure"
  - How is our version of a concept different, why is it?
  - Why is it better for the task we are attempting to augment?
  - Why don't we make a reference to FCA?
  - Why do we seem to be not so concerned with precision?

- Distinction between associative and formal representations
  - How is our system/idea a bridge between those and why is that
    relevant
    - This is the main thrust of Uta's two postings
      - See them, they frame things well, give us entry points
  - Why are associative structures relevant to existing content
    - This provides hook into PORT content
       - Making distinctions between existing content, and
         generated (annotative) content

- Figures
  - Existing figures unclear (visually)
  - Existing figures point unclear
    - What do they represent
    - Need to clarify that they represent clusters before concept
      has been identified: that is, something is there, what is
      it?