QZB
                      design thoughts v.4

It is  not  important  how  quiz-preparing  "infostructure"  is
organized, more important is to have certain functions.

*

The principles

- questions and quizzes must be tagged with classification
data. Such as: 

  * subject       - generally accepted name for the knowledge
  * part          - division of topics convenient for the authors
  *   subpart
  *     ...
  * keywords
  * author
  * compiler      - who collected/combined
  * organization
  * version
  * copyright
  * id            - for unique reference
  * info          - additional info

- content-generators

  * API for task-generators

- questions may contain:

  * formulation (info to be available initially to the student)
  * dialog data, may be:
    * variants of answers
    * possible hints
    * explanations of certain errors
  * verification code (so we know the answer is correct)
  * estimate of task complexity (in scalar or vector form)

- question (task) banks must be manageble

- additional info (usually on quiz-level) may comprise:

  * statistics gathering details, grading at the same time
  * reaction/reply scheme (transition table, cadres), or anyway
  * info for ordering questions
  * modes of operation

- mapping questions to students

  * scheme for mapping

- logging

  * who
  * when
  * which questions
  * how answered

- interapp exchange

Quizzes will not be dependent on certain data formats. QZB must
allow usage of all programs which could do  quizzing.  However,
there must be internal capabilities to perform  some  kinds  of
quizzes. It means, that  QZB  must  allow  generation  of  many
formats of documents once it "does know" the  certain  software
capability  to  handle  quizzing  dialogs.  New  plugin  -  new
function.

It is not necessary to have all the data each time!  Everything
can be omitted. For example, if I  want  just  some  quick  and
dirty quiz done,  I  can  omit  everything  but  questions  and
grading infos and certain defaults will be assumed.