How does a disrupted transposition differ from a plain one?

A plain columnar transposition fills the grid completely, row by row, which leaves regular column lengths that an attacker can anagram. A disrupted, or straddle, transposition leaves an irregular pattern of empty cells on the first pass and fills them later, breaking that regularity.

What does the keyword do?

The keyword's letters are ranked alphabetically to give each column a number. Those numbers control both the disruption triggers as the grid is filled and the order in which columns are read out to form the ciphertext.

How can the receiver reconstruct the empty-cell pattern?

The pattern of empty cells depends only on the keyword and the message length, both of which the receiver knows. So the receiver rebuilds the same grid shape, drops the ciphertext back into the columns and reads the rows in order.

What happens to spaces and punctuation?

Spaces are removed before encrypting so they do not leak word boundaries. Other characters are kept and transposed along with the letters, so strip punctuation yourself first if you want a letters-only message.

Is this cipher secure on its own?

Transposition alone preserves letter frequencies, so it is breakable by analysis even when disrupted. Its real value is as a layer combined with substitution, as in the Soviet VIC cipher, where it greatly increases the work of an attacker.

Straddle Transposition Cipher

The straddle (disrupted) transposition is a columnar transposition cipher with a twist: instead of filling the grid neatly row by row, it leaves an irregular, key-driven pattern of cells empty on a first pass and fills them on a second. This “straddling” of the rows defeats the simple anagramming attacks that break an ordinary columnar transposition, which is why disrupted transposition was a component of strong field ciphers.

How it works

The keyword’s letters are ranked alphabetically (ties broken left to right) to assign each column a number from 1 to N. Those numbers do double duty:

Filling (with disruption). The text is written into a grid row by row, but on each row the cells up to a moving “trigger” column are skipped and left empty; only the cells beyond the trigger are filled. After this triggered first pass lays a diagonal of gaps, a second pass fills every remaining empty cell row by row.
Reading. The columns are then read top to bottom in ascending key-number order, and the characters collected form the ciphertext.

Decryption is exact because the shape of the empty-cell pattern depends only on the keyword and the message length — both known to the receiver. They rebuild the same grid, distribute the ciphertext back into the columns in key order, and read the active cells row by row.

Example

With keyword CONVOY and message RETREAT TO THE EASTERN RIDGE AT ONCE, the spaces are stripped and the irregular grid produces ciphertext such as:

RTENATOSIOETARTEHEGCRTTDNAEREE

Decrypting that string with the same keyword reproduces the original RETREATTOTHEEASTERNRIDGEATONCE.

Notes

Spaces are removed before encrypting so word lengths are not revealed; restore spacing manually after decrypting if needed.
Because transposition only rearranges letters, the ciphertext has the same letter frequencies as the plaintext — pair it with a substitution cipher for real strength.
Disrupted transposition is one of the diffusion stages inside the famous Soviet VIC cipher.