Boak’s Puzzle: Disposing of Classified Trash

Boak's History of US COMSECRecently I was skimming through the NSA’s “classified history of COMSEC” (Volume 1 and Volume 2).  This “history” is a transcription of lectures by David G. Boak, who liked to explain NSA-related topics from a historical perspective. He clearly inspired a generation of NSA’s employees. The last “real” page of the document contains a humorous story and a crypto puzzle (link to extract in pdf).

The NSA had an incinerator in their old Arlington Hall facility that was designed to reduce Top Secret crypto materials and such to ash. Someone discovered that it wasn’t in fact working. Contract disposal trucks had been disposing of this not-quite-sanitized rubish, and officers tracked down a huge pile in a field in Ft. Meyer.

How did they dispose of it? The answer is encrypted in the story’s text!

The story sounds like it’s from the early 1960s. The Arlington Hall incinerator contained a grating that was to keep the documents in the flames until reduced to ash. The grate failed, and “there was no telling how long the condition had persisted before discovery.”

Contract haulers had been dumping the stuff “in abandoned clay pits in Fairfax County” and NSA investigators never found them.

For those of you familiar with Fairfax County today, it may seem hard to believe that there were “abandoned” chunks of land anywhere. I grew up in the wilds of that area back then, and believe me, there were places that seemed pretty remote from civilization.

At any rate, the investigators did manage to track down a couple huge piles of ashes over at Ft. Meyer – about 1,000 cubic yards of the stuff. And here’s the puzzle: the action officer tasked with handling the mess came up with a simple and brilliant strategy – cost effective at the time but probably not any more. What was it?

The Puzzle – Possible Cribs

Boak encrypted the answer in an “innocent text system” in the story’s text itself.I haven’t actually tried to crack the code yet myself. First, I’m racking my brain for strategies of how the disposal might work. Here are some ideas:

  • Dump it in the Potomac – this was back when conservationsts were mostly cranky scientists, agricultural experts, and a few youthful geeks.
  • Dump it in a more salty place – the Chesapeake or the Atlantic.
  • Mix it with cement and use it as a foundation for the Capital Beltway – the incident apparently happened around the time of Beltway construction.
  • Ditto, but pour it into cement foundations for new CIA headquarters being built in nearby McLean, or perhaps for a classified building elsewhere. Ft. Meade, perhaps?
  • Ditto, but use it in construction of the Greenbrier Bunker in West Virginia.

The fundamental problem is to keep control of the readable chunks. While most of the material was ash, there were “palm sized” chunks of waste that might disclose information. Whatever they did, they had to ensure those chunks were unreadable, or at least safe from disclosure.I like the cement idea, actually. No doubt there were cleared construction crews all over the Washington area back then. Temporary structures littered the Mall, most of them built during WWII, and many housed intelligence agencies. They were all gone in the next several years as the occupants moved to newly built, more secure locations.

Some of these construction tasks were sensitive enough to demand some clearances from the construction crews: CIA headquarters, the Greenbrier Bunker, and no doubt Ft. Meade had projects underway, too. Personally, I’m not aware of the construction dates up at Ft. Meade – it wasn’t part of my youthful stomping grounds. But the CIA and the Greenbrier might fit the calendar.

If they used the stuff in the Beltway, they’d need to send officers to ensure that no readable chunks were pulled out by workmen. I doubt any clearance effort went into Beltway construction crews.


You will find the solution in the Comments below. Thanks to Gregory for working it out and checking it.

Note that my cribs were too detailed.


This is a bit like the old

This is a bit like the old Francis Bacon code (different type fonts marked the cipher symbols).

I noticed right off that there’s a misspelling in the first sentence: “baside” for “beside”. Then in line 1 of the last paragraph, there’s “he advised” for “be advised”.

Then the odd sentence in the paragraph before that: “That, right, I have!”

“Our investigator found that site, alright, …” That last is probably an extra word with part of the cipher.

It does go to show that there’s more to this than RSA encryption.

There’s something up with the

There’s something up with the numbers in the text.

The text is very specific about some dimensions (e.g. “five feet in height, eight to ten feet wide” – is there a reason we need to know the exact size?), but vague in other areas (e.g. “some the size of the palm of your hand” – why not use inches?). In the last paragraph, is there any reason for using the numbers “twenty-jillion” and a “million”? He calls special attention to a “million” by adding “a myriad” immediately afterwards, which is incorrect: a myriad is 10,000, not a million.

I noted the spelling errors

I noted the spelling errors as well. The author also, on two occasions, uses digits rather than words for numbers. Perhaps significant. (60’s and 100)

Anyone found the solution?

Anyone found the solution? I’m curious. 🙂

Regarding solutions: I

Regarding solutions: I haven’t heard about any. I plug ‘Boak puzzle’ or some such into Google every once in a while to see if something has popped up elsewhere. Nothing so far.

I think it is a simple

I think it is a simple innocent text system, as advertised, with only one twist that slowed me down. Of course, I may be completely wrong.

I think the solution is: “The telling ash was simple buried” Allowing for one error, as mentioned in the text, this ought to be: “The telling ash was simply buried.”

There are six errors that can be corrected with a letter and six paragraphs on the page. I found that interesting. There was also one transposition error, which has no single unique letter. Hmm, not sure what to do with that – maybe that’s a typographical error (and thus the one deliberate mistake).

In sequence, the errors, and the alphabetic position of the letter are:

1st error: E (5)

2nd error: F (6)

3rd error: N (14)

4th error: E (5)

? error: U/L Transposition

5th error: S (19)

6th error: B (2)

Selecting one word from each paragraph, whose position corresponed to the letter’s alphabetic position, I obtained “The telling ash was with you.” The first four words made me think I had the solution, but the 5th and 6th words don’t make sense.

So I reconsidered the possibility that the transposition error isn’t a deliberate, non-cipher error. I think, instead that it’s an indicator that the transposition code for the system is changed somehow.

Figuring out the shift is a challenge, because there are many possibilities, once the simple alphabet position is dropped; I decided to work backwards a bit. It seems that the only word in the sixth paragraph that is a really good candidate to finish the sentence is “buried.” Nothing in the fifth paragraph leaps out as well, but “simple” seems close to a word that popped into mind immediately after reading the first four words, “simply.”

Wishful thinking, perhaps. But I tried to see what I could work out for the position of these words. If I start counting from the second sentence of these respective paragraphs, the position is 18 for “simple” and 27 for “buried.” If you take the mod(26) operator for these, the two numnbers, 18 and 1, seem suspiciously close to the corresponding letter positions for the errors, 19 and 2, and have the correct spacing.

It is possible to arbitrarily define a shift to the transposition code after the transposition indicator, to make this fit the target text. The shift I posit is:

1. start counting words from the second sentence

2. add one to the basic alphabet cipher (or rotate the transposition so that B = 1)

3. add 26 to the cipher number up to (or through) the letter L

This provides the solution I offered. I think the change after the transposition error is a bit arbitrary, which makes it a little unsatisfying, but if you believe the code is as simple as one word per error, it’s about the only sentence that makes sense.

GregoryF, I believe you have

GregoryF, I believe you have the correct solution. Note the final words of the Boak text: “because that is par for the course.” I think that language is key.

Add up the error positions you describe in your post:

1st error: E (5)

2nd error: F (6)

3rd error: N (14)

4th error: E (5)

? error: U/L Transposition

5th error: S (19)

6th error: B (2)

5 + 6 + 14 + 5 + 19 + 2 = 51,

then, accounting for the “deliberately incorporated error,” add 51, from above, to 19 + 2 = 72, which is, of course, “par for the course.”