requirements for the Key Words in Context System, Initial System
================================================================
assignment 6, Dec. 5, 2002

- required to run on Unix systems only
- runs on standard PC hardware

- all the requirements from the slide

input interface:
- read from file
  - file name: from command line argument
  - contents:
    - 8-bit ASCII
    - line separation: linefeed
    - word separation: blank, tab
    - locale: C
  - size:
    - lines max 1024 chars
    - max 1000000 lines
    - max 10000000 chars
    - may use temporary files
- sorting:
  - by ASCII order
- output:
  - write to file
    - name: from command line argument
  - just the shifted, sorted lines
  - 8-bit ASCII
  - line separation: linefeed
  - word separation: blank
- command line arguments
  - two, both mandatory
  - first: input file
  - second: output file
- should always terminate if the other requirements are met
- some "reasonable" time for processing

modules and their interfaces
============================
general: use C++ as implementation language

- line holder module
  - exported data types:
    - text
    - line
    - word
    - character
  - interface programs:
    (for the parameters: all counts start at 0)
    - inline void text::setChar(unsigned int lineNo, unsigned int wordNo,
              unsigned int charNo, character ch)
    - inline character text::getChar(unsigned int lineNo, unsigned int wordNo,
              unsigned int charNo)
    - inline unsigned int text::lines()
    - inline unsigned int text::words(unsigned int lineNo)
    - inline unsigned int text::characters(unsigned int lineNo,
              unsigned int wordNo)
- input module
  - interface programs:
    - text *input(const char *inFileName)
- circular shifter module
  - interface programs:
    - class circularShift: public text
    - circularShift::circularShift(text &inText)
- sorting module
  - interface programs:
    - class alphSort: public text
    - alphSort::alphSort(text &inText)
- output module
  - interface programs:
    void output(const char *outFileName text &inText)
- master control module
  - interface programs:
    - int main(int argc, char *argv[])
- command line module
  - interface programs:
    - cmdLine::cmdLine(int argc, char *argv[])
    - const char *cmdLine::inFileName()
    - const char *cmdLine::outFileName()

open problems:
==============
- Are multiple, immediately consecutive word breaks preserved?
  -> proposal: no
- How do word breaks sort?
  -> proposal: like a single blank
- We don't need the data types "line" and "word" in this setup.
  -> proposal: delete them