Histogram generation

Python - 83 characters

Seems that we can take input from anywhere, so this takes input during execution, rather than from the command line, and uses Ejrb's suggestion to shorten it by 8.

s=map(len,raw_input().split())
c=0;exec'c+=1;print"%3d|"%c+"#"*s.count(c);'*max(s)

Python - 91 characters

This will fall over with quotes.

import sys;s=map(len,sys.argv[1:])
for i in range(1,max(s)+1):print"%3d|"%i+'#'*s.count(i)

Input:

> python hist.py Please write a specification rather than giving a single example which, solely by virtue of being a single example, cannot express the range of acceptable output styles, and which doesnt guarantee to cover all corner cases. Its good to have a few test cases, but its even more important to have a good spec.

Output:

  1|#####
  2|######
  3|#####
  4|##########
  5|######
  6|#############
  7|####
  8|#
  9|##
 10|#
 11|
 12|
 13|#

R, 55 47 characters

hist(a<-sapply(scan(,""),nchar),br=.5+0:max(a))

Luckily R comes with a plot function hist for histograms, here supplied with a breaks argument where the breaks are 0.5, 1.5, ... until max(input)+0.5. sapply(scan(,""),nchar) takes an input (as stdin), separates it following the spaces and count the number of characters of each element.

Examples:

hist(a<-sapply(scan(,""),nchar),br=.5+0:max(a))
1: Extensive word length should not be very problematic.
9: 
Read 8 items

enter image description here

hist(a<-sapply(scan(,""),nchar),br=.5+0:max(a))
1: Very long strings of words should be just as easy to generate a histogram just as short strings of words are easy to generate a histogram for.
28: 
Read 27 items

enter image description here

Edit:

A variation at 71 characters with an axis label at each possible value:

hist(a<-sapply(scan(,""),nchar),br=.5+0:max(a),ax=F);axis(1,at=1:max(a))

enter image description here


Haskell - 126 characters

p[d]=[' ',d];p n=n
h l=[1..maximum l]>>= \i->p(show i)++'|':(l>>=($"#").drop.abs.(i-))++"\n"
main=interact$h.map length.words

This takes the input from stdin, not the command line:

& head -500 /usr/share/dict/words | runhaskell 15791-Histogram.hs 
 1|##
 2|##
 3|######
 4|###############
 5|################################################
 6|###############################################################
 7|###################################################################
 8|###########################################################################
 9|#############################################################
10|##########################################################
11|#########################################################
12|#########################
13|#######
14|###
15|#####
16|###
17|#
18|
19|#
20|#