# Last edited on 2003-12-31 16:33:01 by stolfi
# Analyzing Reid's First Language List

INTRO

  This directory records an attempt to analyze the evolution of 
  Richard J. Reid's list of computer languages used in introductory
  programming courses.  The main analysis was done in Sep. 2001,
  with a small update in Dec. 2003.

DATA

  For the Sep/2001 summary, the raw lists were obtained from the
  postings by Richard Reid himself. Currently all those lists, plus a
  couple of new ones, became available at

  http://www.csee.wvu.edu/~vanscoy/reid.htm
  
DATA CLEANUP

  The following list versions were considered (the last one in 
  the Dec/2003 update only):
  
    set vers = ( 1995-04 1999-08 2001-02 2002-02 )
  
  The raw lists were cleaned up in several ways, resuling in files
  "{VERS}-all.tbl", with the following format
  
    LANG FRAC VARIANT {SCHOOL}
    
  where the fields are
  
    LANG    generic language (E.g. "Pascal", "Modula");
    FRAC    fraction of time/students devoted to LANG;
    VARIANT "-" or variant of LANG (e.g. "ObjPascal", "Modula-2");
    SCHOOL  school's name.
   
  Schools that were listed by Reid as using 2 or more languages were
  split into separate records with the appropriate {FRAC} fields. The
  generic language names were inserted by hand.
  
  For 2002-02, three lists were generated
   
    2002-02-cfm.tbl   Data updated or confirmed between 2001 and 2002
    2002-02-unc.tbl   Data from 2001 that was not confirmed/updated
    2002-02-all.tbl   Union of the above.
  
  The scripts {extract-univs} was used to extract the university
  names:
  
    foreach f ( ${vers} )
      echo $f
      cat ${f}-all.tbl | extract-univs | sort -u > .uns-$f
    end
    
    cat .uns-????-?? | sort -u > .uns-all
  
  The script {fix-school-names} was generated by amnually editing this
  list, and was then applied to each file, so as to ensure that the same
  spelling and form of the name was used in all lists
    
  Checking that "new"schools are not typos:
  
    cat .uns-{1995-04,1999-08,2001-02} | sort -u > .uns-old
    bool 1-2 .uns-2002-02 .uns-old
    
  The output must be compared manually to .uns-old, looking for 
  near-duplicates.  Beware that "Brunei U." and "Brunel U." both exist,
  and ditto for "DePaul U." and "DePauw U.".

TALLYING

  Tallying the tabulated data:

    foreach f ( ${vers} )
      echo ${f}-all.tbl
      cat ${f}-all.tbl | tally-langs | sort -b +0 -1 +2 -3gr > ${f}-all.tally
    end

    cat 2002-02-cfm.tbl | tally-langs | sort -b +0 -1 +2 -3gr > 2002-02-cfm.tally
    cat 2002-02-unc.tbl | tally-langs | sort -b +0 -1 +2 -3gr > 2002-02-unc.tally

    1995-04-all.tally
    1999-08-all.tally
    2001-02-all.tally
    2002-02-all.tally


    2001-02-common.tbl
    1995-04-common.tbl
    1999-08-common.tbl
    2001-02-common.tally
    1995-04-common.tally
    1999-08-common.tally
    
  
  
  .