[ucc] Progcomp results

[Grahame Bowland]

Hi all,

The winners of the 2002 Programming Competition are:

Q1: Nick Bannon
Q2: Michael Pauley
Q3: Rodney Lorrimar

Apologies for the long delay. It was a combination of a broken 
MacOS install, not having my <expletive> together and trying 
to fairly judge Q1. Attached are my notes.

The level of entries was amazingly good! Thanks to everyone 
for participating.

Have fun
grahame

-------------- next part --------------

Entries are judged on the following criteria:

1. For their output, are they capable of decrypting to produce the original 
   cleartext? Is this an implementation of solitaire?
2. Other discriminating factors?

Note: juding this question was made much harder by the fact that no standard 
      way of defining the inputs to the program was specified in the question. 
	  As a result each program was tested with a variety of keys generated 
	  for its particular input specification, with a common test phrase.

Nick Bannon
-----------

1. * code compiles (Linux/2.4.18/i386)
   * code runs
   * for all five tested key sets, it produced satisfactory output 
	 and could decrypt the output into the original cleartext
   * stress-test: large raw file. Handles case well with good memory 
	 usage.
2. * useful key generating utility 
   * written in relatively portable C code

Thomas Catiglione
-----------------

Tested on MacOS X 10.2 (Jaguar) with latest compiler environment. The judge 
had some problems getting this to link, which turned out to be an old version 
of gcc - no fault of entrant.

1. * code compiles (Darwin/10.2/PowerPC)
   * code runs
   * for all five tested key sets, it produced satisfactory output
	 and could decrypt the output into the original cleartext
   * stress-test: large raw file. Handles case badly, had to force-quit. A quick 
	 look in top showed the application rapidly allocating large amounts of memory, 
	 when killed it had reached 200Mb allocated.
2. * well written Aqua GUI
   * nice display of the state / generation and display of the keystream

Trent Lloyd
-----------

1. * code compiles (Linux/2.4.18/i386)
   * code segfaults
2. * the actual implementation of solitaire is in there and looks right, it just 
     hasn't been finished off to actually read in a key / plaintext and output it

Conclusion
----------

Therefore the winner of Question 1 is Nick Bannon, as his implementation is both 
correct and handled the edge cases well. Runner up is Thomas with his very nice 
GUI implementation.

-------------- next part --------------

Nick Bannon
-----------

 * runs (perl 5.6.1 for i386-linux)
 * perhaps will never terminate on 32 bit architectures as I suspect we'll overflow first *grin*
   unless perl has some clever way of fixing this
 * cute use of figlet :-)

Rodney Lorrimar
---------------

(Author only has a K6/2 without acceleration under X at home.)

 * this kicks ass
 * really cute, if amazingly slow on my computer (no fault of author)
   I may get addicted to this at work

Conclusion
----------

Winner of Q3 is Rodney Lorrimar. Nick's entry also cute.

-------------- next part --------------

Michael Pauley
--------------

1. Program compiles (Linux/2.4.18/i386)
2. When given the UCC homepage to parse, the program outputs suitable 
   output
3. Examination of the code indicates that it is doing depth counting, 
   that it is suitably parsing XML. Does not appear to handle <li/> 
   elements though.

Conclusion
----------

The winner of Question 2 is Michael Pauley