icde-related-work

ICDE Related Work

Use this to audit novelty and eligibility. Reopen the current call for the original-work and

prior-publication rules before advising authors.

Positioning checks

• Separate genuine novelty (a new mechanism, access method, protocol, or primitive) from

engineering delta (a faster implementation of a known idea). ICDE rewards the former;

the latter needs an unusually strong evaluation to survive.

• Run a reinvention audit against the deep database canon. Because data management has a

fifty-year literature, a "new" index or join often has a 1980s or 1990s ancestor; find it and

state the difference, or a reviewer will.

• Cover both the research venues and shipping systems. ICDE reviewers expect you to know

the relevant production databases, not just papers; ignoring a system that already does what

you claim is a recognizable reject pattern.

• Because ICDE is single-blind, cite your own prior work directly and link repositories

freely — there is no anonymity to preserve in the bibliography.

• State overlap with any prior short/workshop version honestly; the original-work rule bars work

already published beyond a few double-column pages.

The venue-ring coverage table

| Literature lane | Typical sources | What ICDE reviewers check |

|---|---|---|

| Core database venues | ICDE, SIGMOD, VLDB/PVLDB, EDBT | Whether the nearest system is compared or explicitly distinguished |

| Journals | IEEE TKDE, The VLDB Journal, ACM TODS | Whether the journal-length predecessor of your idea is acknowledged |

| Systems venues | OSDI, SOSP, FAST, ATC | Whether a systems-community mechanism you rebuild is credited |

| Shipping systems | Commercial and open-source databases | Whether a production system already provides your primitive |

A bibliography citing only recent conference papers tells a builder-heavy committee that a

known technique may be getting rediscovered — no throughput number repairs that impression.

The concurrency question at ICDE

• ICDE's two-round calendar and the tight ICDE/SIGMOD/VLDB overlap mean genuinely concurrent

work is common. For independently concurrent papers, cite neutrally, state the technical

difference, and avoid priority claims a reviewer cannot check.

• If a near-identical result appeared at a sibling venue between your rounds, address it head-on

rather than hoping the reviewers missed it — they did not.

Positioning vignette

A paper proposes a latch-free variant of a classic index. Its nearest neighbors: the original

B-tree literature, a recent SIGMOD lock-free structure, and a production key-value store's

internal index. The novelty sentence must name all three contrasts — a new mechanism the

classic design lacked, a different hardware assumption from the SIGMOD variant, and a capability

the production store does not expose — so the delta is unmistakable.

Overlap judgment calls

• Your own workshop or short-paper version: usually citable, but confirm it stays under the

original-work threshold and declare it.

• When unsure whether a prior venue counts as prior publication, declare the overlap in the

submission form rather than gambling on a chair's reading.

Output format

[Eligibility] clear / needs declaration / risky
[Novelty type] new-mechanism / engineering-delta
[Closest lanes] <core-DB / journal / systems / shipping-system>
[Nearest 3 works] <work -> specific distinction>
[Reinvention risk] <ancestor found? y/n>
[Novelty sentence] <ICDE-ready contribution contrast>