Grossi–Vitter constant-time suffix-array representation question

Does there exist a family of data structures which, for every binary text T{0,1}nT\in\{0,1\}^n, uses O(n)O(n) bits of space and answers suffix-array access queries SAT[i]\operatorname{SA}_T[i] for every i{1,,n}i\in\{1,\ldots,n\} in O(1)O(1) cell probes, in the cell-probe model with word size Θ(logn)\Theta(\log n)? Here SAT[i]\operatorname{SA}_T[i] is the starting position of the ii-th suffix of TT in lexicographic order.

Sources & referencesView supporting material

Progress summary

Refreshed
Claimed solved

A new preprint claims to rule out constant-time suffix-array access using only linear space on binary texts, but the claim has not yet been independently verified.

The question asks whether suffix-array access can be represented in linear space with constant query time, even for binary texts. A new preprint claims a cell-probe lower bound that rules this out.

August 2026 cell-probe lower bound

The preprint claims a nontrivial time-space lower bound and complexity-preserving reductions to prefix-query problems; its stated consequence is that linear-space, constant-time suffix-array access is impossible for binary texts. Earlier work established bidirectional reductions between suffix-array queries and prefix-select queries, together with subpolynomial-time tradeoffs, but did not establish this impossibility result.

Current status (as of August 2026): A preprint claims the lower bound settles the impossibility direction, while independent verification and peer-reviewed confirmation are not recorded.

Sources

Solutions 0

No solutions have been posted yet.