We study the compressibility of enumerations, and its role in the relative Kolmogorov complexity of computably enumerable sets, with respect to density. With repspect to a strong and a weak form of compression, we examine the gain: the amount of auxiliary information embedded in the compressed enumeration. Strong compression and weak gainless compression is shown for any computably enumerable set, and a positional game is studied toward understanding strong gainless compression.
Let $C(\cdot|\cdot)$ denote the conditional Komogorov complexity.
Let $A\subset\omega$. The first $n$ bits of $A$, denoted by $A\upharpoonright_n$, is the $n$-bit binary string $\chi_A(0)\chi_A(1)\cdots\chi_A(n-1)$.
Let $A,B\subset\omega$. $A$ is rK-reducible to $B$, written $A\leq_{rK} B$, if $C(A\upharpoonright_n|B\upharpoonright_n)\leq O(1)$. Equivalently, if there is a constant $k$ and a partial computable function $f$, such that $f(B\upharpoonright_n)$ outputs an index $e$ such that $A\upharpoonright_n\in W_e$ and $|W_e|\leq k$, where $W_e$ is the $e^\text{th}$ r.e. set.
Given r.e. set $A$, we call a r.e. set $D$ a compression of $A$ if $A\leq_{rK} D$.
A set $A$ is well-compressible if it has a gainless strong compression.
Theorem. Each rK-complete set (that is, an r.e. set $K$ such that $A\leq_{rK}K$ for any r.e. set $A$) has a strong gainless compression.
Theorem. Each r.e. set has a strong compression.
Theorem. Each r.e. set has a weak gainless compression.
Theorem. Let $r$ represents any of $rK$ reduction, $K$ reduction or $C$ reduction. If r.e. sets $A,B$ both have strong gainless compression and $B<_r A$, then there is an r.e. set $C$ such that $B<_r C<_r A$.
Does each r.e. set have a strong gainless compression?