How to crack the Vigenère cipher if the source text is a random set of characters?

Question

Encryption

Suppose there is a source text, which is a sequence of absolutely random characters. And there is a keyword, which is already a meaningful set of characters (some simple word).
The text is then encrypted with the key.
For example, the situation is:
text : "lkh*AS&)(*#opsJFsfa9847ksydfO^&ASD(*(XU9a8x3yj"
keyword : "simple"
key: "simplesimplesimplesimplesimplesimplesimplesimplesimple"
I'm with cryptography on "You" or even on "Who are you?". From one publication on Habré was able to make the following:
- Vigenère decryption is based on determining the key length;
- the length of the key gives out that the same characters located in the source text at a distance that is a multiple of the length of the key are encrypted in the same way, that is, repeated sequences are searched;
- further it is known that the characters located at a distance from each other, a multiple of the length of the key, are encrypted with one cipher and:

This makes it possible to use the known characteristics of the frequency distributions of plaintext letters to break each monoalphabetic cipher separately.

- this is the moment that interests me. Do I understand correctly that this stage of decoding implies that the source text contains language constructs (or even consists only of them). If so, what can be said about a sequence of random characters encrypted with a Vigenère cipher using a simple word as the key? How difficult would it be to decipher this?
On an intuitive level, this seems to me indecipherable because the mere assumptions about the source text, when it is absolutely random, seem to me absurd. But perhaps encryption will introduce some pattern?
Help me to understand.

Answer 1

If you swap text and key, the cryptogram will not change. However, it will no longer be Vigenère, but a one-time pad that has absolute durability. Absolute security does not imply an infinite number of possible keys, but the impossibility of checking the correctness of the decrypted result.
upd: However, this is only true if the text sequence is completely random. But generating such sequences is practically very difficult. Therefore, they resort to an attack on the sequence generator itself. Most of them have seed values ​​that you can try to pick up.