No. You mean you need at least that to represent any such message in a constant-size format. That’s not what the complexity is about—if the message is the alphabet and then the rest are spaces, that’s the right length and has the right number of symbols, but you can easily compress it to much much less than 381 bytes.
That said, I agree that Java bytecode probably isn’t the ideal medium for transmitting such a terse message.
I think past!me miscommunicated and was stupid w.r.t. Kolmogorov Complexity, and I think you have misread the statement too.
The program “enumerate all X symbol strings from this Y letter alphabet then choose the Nth” is pretty much one of the simplest ways of encoding a string. So I was merely remarking that to single out one 80 symbol, 27 letter alphabet string, you need at least 381 bits, or you will have incomplete domain coverage due to the pigeon-hole principle.
If we expand the alphabet to 32 letters, then it obviously takes 400 bits, which is 50 bytes.
The program “enumerate all X symbol strings from this Y letter alphabet then choose the Nth” is pretty much one of the simplest ways of encoding a string.
A completely arbitrary string, yes. This was nowhere near an arbitrary string.
No. You mean you need at least that to represent any such message in a constant-size format. That’s not what the complexity is about—if the message is the alphabet and then the rest are spaces, that’s the right length and has the right number of symbols, but you can easily compress it to much much less than 381 bytes.
That said, I agree that Java bytecode probably isn’t the ideal medium for transmitting such a terse message.
I think past!me miscommunicated and was stupid w.r.t. Kolmogorov Complexity, and I think you have misread the statement too. The program “enumerate all X symbol strings from this Y letter alphabet then choose the Nth” is pretty much one of the simplest ways of encoding a string. So I was merely remarking that to single out one 80 symbol, 27 letter alphabet string, you need at least 381 bits, or you will have incomplete domain coverage due to the pigeon-hole principle.
If we expand the alphabet to 32 letters, then it obviously takes 400 bits, which is 50 bytes.
A completely arbitrary string, yes. This was nowhere near an arbitrary string.