This evening I submitted the final exam for the third of the three papers I'm taking this semester. I now have about a three week break to do whatever I want. I'll have to rearrange the second semester's papers in that time.

The exam is online, and you're allowed to look things up on the internet during it. I answered a couple of the questions by typing a keyword into DuckDuckGo and looking at the Wikipedia preview box.

I rediscovered a website that I lost the URL to more than a year ago. It's for performing frequency analysis on text, which helps to break some relatively simple text coding algorithms designed for humans, like the Caesar cipher, but also ones which shuffle the output instead of shifting it. The tool is interactive and helps you crack it letter by letter. It's really cool.

Try it yourself, it's fun! Here's an encoded message:

J'E TNB INNXJTG HZ QTVBWJTG NTIJTM, J'E BPVJTG BN RJTF QTF RJGHPM NHB CBPQBMGJMC QTF CWNPBOHBC EVCMIR BWQB QPM KJBWJT EV CXJII IMLMI. J'E FNJTG BWJC KJBW Q RMK ZMNZIM NTIJTM, QTF J'E CWQPJTG EV BJEMC QTF CBPQBMGJMC KJBW BWME, QTF BWMV QPM CWQPJTG YQOX. BWMV BNIF EM QYNHB Q DHEZ JT EJPPNP BMEZIM BWQB IMBC VNH CXJZ GMBBJTG BKN XMVC. JB'C PMQIIV MQCV BN FN QTF CQLMC BKN EJTHBMC.

All letters are changed to capital letters, but puncutation is preserved.

To get started, go to the frequency analysis website, paste that code into the top most box, and press "decrypt automatically". The website will calculate which letters appear most often, and compare those frequencies to the words that appear most often in the English language. There are only a couple of sentences, so there's not enough data for it to get a perfect translation. I tried this on longer paragraphs, and often it works perfectly.

The top line is the original text, and the bottom line is what it thinks is the correct translation.

The initial decryption won't be very good, and you probably won't be able to make out many of the words yet, but quite a few of the letters are actually correct. The hard part is figuring out which ones are correct and which ones need to be changed.

If you're sure that a letter should be a different letter, then click it on the bottom row, press Ctrl-A to select it, then type the correct letter to override it. The letter will change everywhere. This allows you to interactively decode the message as you figure out more and more parts. Cool, isn't it!

If you're up for the challenge, try translating that code block above. The original message is adapted from a few sentences that I posted on this blog a few days ago. The post in question does not contain any computer jargon. I left in the punctuation and letter spaces to help you. Go! Try it! Now! See if you can work out what it says!

If you need some help, then click these boxes to progressively reveal hints...

Hint 1: Look at the many words that were originally J'E. Remember, the apostrophe is correct. What could these be?

Same hint continued: Can you think of any words, that would fit in a personal blog context, that are two letters with an apostrophe in between?

Answer to that hint: J'E translates to i'm.

Hint 2: Look at JB'C. You already discovered that J translates to i from the previous hint, so you're trying to decode i?'?. What could that be?

Answer to that hint: JB'C translates to it's.

Hint 3: What about that lone word Q? What words only have one letter in them?

Same hint continued: You already found out that J translates to i in hint 1, so Q can't also translate to i. It must be a different letter.

Hint 4: With punctuation done, there are a few good options for regular words that are now revealed. Can you think of any replacements for LIVIL (formerly IMLMI), USIOR (formerly FNJTG), or TDIS (formerly BWJC)?

Answer to that hint: The translations for those words are level, using, and this.

From here on, most words should be almost finished, and some entirely finished. Look for parts of sentences and use those parts to help you guess what one word might say.

Good luck!

— Cadence