Icalp topic selection

ICALP Topic Selection

Decide the venue and the track before drafting. ICALP — the EATCS International Colloquium on Automata, Languages, and Programming — is Europe's flagship theoretical-computer-science conference and the annual meeting of the European Association for Theoretical Computer Science. Its contribution is a theorem: a precise statement in a clean model, proved completely. A paper whose real deliverable is a system, a benchmark result, or an application with no new theorem is respected and then rejected as out of scope.

Two decisions, not one

Routing to ICALP is genuinely two questions:

1. Is this an ICALP paper at all, versus STOC/FOCS/SODA, a specialized venue (LICS, CCC, SoCG, CONCUR, ...), or a journal?
2. Which track — Track A (Algorithms, Complexity and Games) or Track B (Automata, Logic, Semantics and Theory of Programming)?

Both matter: the tracks have separate program committees and separate HotCRP servers, and Track B carries a rebuttal while Track A does not.

Track A vs Track B

Signal in your result	Track
A faster algorithm, a better approximation ratio, a new data structure, a fine-grained bound	A
A complexity lower bound, hardness, or a completeness result about computation	A
Algorithmic game theory, mechanism design, equilibrium computation	A
Decidability/complexity of a logic or a theory, model theory of computation	B
Automata, formal languages, transducers, infinite-state systems (Petri nets/VASS, pushdown)	B
Semantics of programming languages, lambda calculi, type theory, verification foundations	B
Concurrency theory, process calculi, category-theoretic semantics	B

Borderline cases exist (e.g. logic-flavored complexity). Resolve them by which community should referee the proof — the one that will recognize the technique and the prior bound — and by which track's recent programs contain your subarea.

ICALP vs the US theory venues and journals

Signal in your project	Better home	Why
Broad TCS result, ready now, and the February ICALP deadline is the nearest honest fit	ICALP	EATCS European flagship; LIPIcs open access; two tracks
Top-tier breakthrough aimed at the US theory community, STOC/FOCS timing fits	STOC / FOCS	ACM/IEEE flagships; different calendar, single-column ACM/IEEE style
Core is discrete algorithms / data structures with a SODA fit	SODA	SIAM algorithms venue; different community and calendar
Logic in computer science specifically (not general Track B breadth)	LICS	Dedicated logic venue; overlaps Track B, different community
Computational complexity specifically	CCC	Dedicated complexity venue
Computational geometry	SoCG	Dedicated geometry venue
Result too long, or wants a definitive journal version of record	journal (JACM, TALG, TCS, LMCS, ...)	No conference page ceiling; some ICALP papers later have journal versions

ICALP's distinctive pull is European TCS community + LIPIcs open access + a two-track single February cycle. It is not STOC/FOCS/SODA; do not carry their templates, blinding regimes, or calendars across.

Contribution shapes ICALP rewards

• A better bound — a faster algorithm, a smaller approximation ratio, a tighter complexity bound than the best known, with the improvement stated against a named prior result.
• A lower bound / hardness / impossibility — proving a limitation, conditionally or unconditionally, in a precisely defined model.
• A decidability / complexity classification — pinning the exact complexity of deciding a logic, language, or verification problem (a signature Track B move).
• A dichotomy or complete characterization — classifying an entire family rather than one instance.
• A new model or a foundational result — a clean model with a clear TCS payoff; "foundations" is in the name for both tracks.

The re-route tests

• Theorem test: strip the implementation and the experiments — is there still a theorem worth proving? If not, the paper belongs at an experimental-algorithms, database, or systems venue, not ICALP.
• Prior-bound test: can you name the best previous bound your result beats? If you cannot, either the problem is not yet a recognized ICALP problem, or the contribution is not a bound (an icalp-writing-style framing fix, or a re-route).
• Track-referee test: who should check the proof — an algorithms/complexity person (A) or an automata/logic/semantics person (B)? If the honest answer is "a systems/ML/database person," ICALP is the wrong venue.

Cheap reconnaissance before committing

[Scope]    scan the last two ICALP programs (dblp db/conf/icalp) for your subarea and track
           -> present in Track A/B recently = a reviewer pool exists; absent = opening or mismatch
[Citations] is your bibliography majority TCS venues (ICALP/STOC/FOCS/SODA/LICS/CCC + TCS journals)?
           -> majority applied venues => reviewers read you as a visitor; reconsider the venue
[Calendar] ICALP has ONE annual February deadline; compare with STOC/FOCS/SODA/LICS dates
           -> route to the nearest honest fit rather than waiting a full year for a marginal preference

Decision procedure

[Is there a theorem?] strip code/experiments -> new theorem remains? no => not ICALP
[ICALP vs US/journal] European TCS breadth, Feb deadline nearest, open-access wanted? -> ICALP
[Track] bound/algorithm/complexity/games => A ; logic/automata/semantics/verification => B
[Specialized check] pure logic => LICS? pure complexity => CCC? geometry => SoCG?
[Verdict] ICALP Track A / ICALP Track B / sibling venue / journal, with a one-line reason

Run this before the writing skills; a wrong venue or track decision wastes every later step. When the verdict is ICALP, continue with icalp-workflow for the calendar and icalp-writing-style for the paper shape.