Ian Bicking: the old part of his blog

Even more significant whitespace

Inspired by the Whitespace language, I thought this should be a more general tool. Not just whitespace programs, we should have whitespace everything. So I did it the cheap way, and just encoded ASCII to whitespace: whitespace.py

Mortgage Quote Request Approved. Four lenders have approved you for the nations lowest rates of under 3%. Please continue to lock in your rate and start saving. Useful as a secret code. Not as good as quantum encryption, this will only detect a small set of interference. All your messages will be invisible to Outlook users. Since the messages are invisible, one would think blind users would do better, but I think they'll be even more confused. Anyway, here's an example message:

Mortgage Quote 		     	 	  		    		    		     	 
Request Approved.		  	 		

lenders	have approved	
you for	
low rates (3%!!!)		
(see how secretive
this can,be?)				


click http://blahblah.com to unsubscribe(or	not)	     	 	   	 
blah blah,you could obviously	write more	  		    		    		     	 	  		     	 
Created 21 Oct '04
Modified 14 Dec '04


Very interesting application, Ian.

I'd love to read your sample message, but whitespace.py bails out after trying to decode the second 6 character whitespace sequence in your message, '\t\t\t \n ', which decodes to ASCII code 286 and thus ord(code) throws a ValueError (which the try..except should prolly catch as well).

The first sequence of 6 whitespace characters (' \t\t\t\t\t') is decodable, it is the letter 'd'.

Should I assume your secret message to be compromised and destroy the plans for the secret project you tried to communicate with me about?


Agent X
# Agent X


I had to do this to get it to work though: note that whitespaceO.py is the original file.

--- whitespace.py Fri Oct 22 21:25:38 2004
+++ whitespaceO.py Fri Oct 22 21:24:32 2004
@@ -40,7 +40,7 @@

def enc_stream(input, output):
while 1:
- s = input.readline()
+ s = input.input.read(4096)
if not s:
for c in s:
@@ -49,7 +49,7 @@
def dec_stream(input, output):
s = ''
while 1:
- data = input.readline()
+ data = input.read(36)
if not data:
s += data
@@ -81,7 +81,7 @@
if args[1:]:
input = open(args[1])
if args[2:]:
- output = open(args[2], 'w')
+ input = open(args[2], 'w')
if use_repr:
stream_output = StringIO()

I wonder if it's a Windows issue? I haven't tested there. Obviously line-ending issues are very important.
# Ian Bicking

I would suggest opening both files in binary mode, so that nobody interprets your line endings for you. ;) Then it shouldn't matter what OS you are on.
# Shane