HACKvent 2015: Day 11

Challenge

The following image is provided:

Solution

I had to spend some time researching the above image but I soon discover that it is a punch card.
It turns out to be a IBM 96 Column Punchcard. Unfortunately there isn't much information online on how to decode them.

A reverse image search comes up with the following image:

This is what the challenge organiser used to make the above challenge!
The first thing I do is open the image in Photoshop and apply a black background so I can see the data (black dots) more easily.

I get:

Another important piece of information was this image I found online:

I used the above image to determine how I decoded the data. The above image encodes the data ZORCH IT'S A SYSTEM 3.

You have to look at each column in the punch card and generate a number.
For example, in the first column above we have _A__81 or in binary 011001.
This corresponds to the letter Z (more on this later).

Thus each block can have a total of 32 bit characters. The total card can therefore contain a total of 96 (32*3) characters.

Now, going back to our message, I first write out all the bits in one big block, I get:

00110100110001111100110111101100 01000110111110111011100111101100 00111101110110110010111101011100
01011101110001001010101100111010 01100100101110110111101001011000 01001001010001100100111000111110
00110000100000000000000010000010 01000100101000110000110001010000 00000100000000010000001000000000
01000100010100010000000101001000 00101110011011010111101111001000 01111000110011100010100101111100
01001001000101111110011000010100 00000101001110011010101101101100 00110000000111000110100000011010
00111101010000101000111111111100 00111101011011011001100101010000 00111101100100110100111101011110

At this stage all I need to figure out is the language, I don't find much on this online but I eventually come up with a 64 length partially working language (it begins with a space):

 123456789????????STUVWXYZ???????JKLMNOPQR???????ABCDEFGHI?????|

So in the example I discussed, 011001 is Z because 011001 is 25 in decimal and Z is at index 25 in the string array above.

Note: This language was created using trial and error and so I didn't determine many characters initially (which I have just left as a ?). Later on I discovered an image (link) which allowed me to determine every character! (some special characters were still left as a ?).

At this stage I can write a nice short python script to read the data and convert it to the language.

# IBM 96 Column Punch Card decoder (Based on EBCDIC)
# Source of information: https://en.wikipedia.org/wiki/EBCDIC

# Language decuded from: http://www.quadibloc.com/comp/images/cardtabl.gif (TOP RIGHT)
language = " 123456789:#@'=?0/STUVWXYZ&,%_>?-JKLMNOPQR!$*);?.ABCDEFGHI?.<(+|"
rows = []
codewords = []

with open('input.txt') as f:
  for line in f:
    list = line.rstrip().replace(' ', '') #ignore whitespace
    rows.append(list)

# make codewords
count = 0
while count < len(rows[0]):
  tempcode = ''
  for row in rows:
    tempcode += row[count]
  
  codewords.append(int(tempcode, 2))
  
  count += 1
  
# Print out codeword
count = 0

for code in codewords:
  print language[code],
  count += 1
  
  # Seperate into 32 blocks
  if count == 32:
    count = 0
    print '\n',

I put the bit data into a file called input.txt and run my script.
The following message is printed out:

  W R I T E   T H E   6   B L O C K S   A L T E R N A T E L Y
  H V 1 5 | M 3 H N | B G 5 H | L U F E | 8 W P M | K Z F K
  U P P E R   A N D   L O W E R   T O   G A I N   N U G G E T

Great! The clue is quite simple, we simply need to alternate the case of each of the 6 blocks in the nugget (as separated with |). We know the nugget begins with the uppercase HV15 so we know the pattern straight away.

Flag:

HV15-m3hn-BG5H-lufe-8WPM-kzfk