* Test constant 'qr'

* 'true' and 'false' properties are opposites that still only affect
  one hash key; define other opposites as well?

* test what happens to propagated properties when you delete them
  on the original object

* overload '0+' and the other missing ones

* in value(), also check for other escaped chars like '\t'

* use it for versioning so you can undo changes? or as in debugging,
  see the whole history of commands that affected a variable.
  Cross-references to other variables that were also affected in
  those statements. A graph...?

* in a similar vein, methods like C<length> could add a property
  to the resulting value that notes where the value comes from:

	my $l = 'hello world'->length;
	print $l->origin;     # prints "length 'hello world'" (or so)

* group methods and make each group a mixin class in a separate
  module, loaded on import. ':all' loads all; 'split' loads only the
  split method; ':num' loads all numeric methods etc.
  The versioning idea above would certainly be a mixin.
  The propagation methods that are automatically exported so far also
  should go into an import tag (':pass'?)

* methods that generate an array can be lazy; in this case they
  return an iterator instead of the list. Cf. mjd's talk.

* take ideas from Ruby's Numeric and String classes, and from Java's
  classes. New methods:
  
  chr   # string containing the ASCII character represented by the value
  ceil  # Returns the smallest integer value greater than or equal to the value
  floor # Returns the largest integer value less than or equal to the value
  round # Returns the value rounded to the nearest integer

  utf8 upgrade, downgrade
  crlf conversion
  printf
  sprintf
  substr
  sub
  	$a = "hello there";
  	$a->sub(1)  === 'e'
  	$a->sub(1..3) === 'ell'
  	$a->sub('lo') === 'lo'
  	$a->sub('bye') === undef
  	$a->sub(/th[aeiou]/) === 'the'
  	this returns a (builtin) substr, which is lvaluable, so you can do
  	$a->sub(/[el]+/) = 'xyz'   # 'hxyzlo there'
  match
  subst (regex subst)
  tr (y)
  charlist # split //
  chomp
  chop
  crypt($salt)
  center($len)
  ljust($len)
  rjust($len)
  trim, ltrim, rtrim
  dump # replace all nonprinting chars by \nnn, escape special chars
  each_byte # list of all ords
  empty # true if length is 0
  index
  rindex
  succ (next, incr) # overloads ++
  pred (prev, decr) # overloads --
  replace # straightforward replace, no regex chars
  squeeze # 'hello world' === 'helowrd'
  md5 # require Digest::MD5
  sum($n) # sum of binary value of each char modulo 2**$n - 1, a weak checksum
  unpack
  upto  # 'a'->upto('e') = qw/a b c d e/;
  clone (deep copy, needs Clone module, so definitely as a mixin class)