Implementing Grassroots Logic Programs with Multiagent Transition Systems and AI (Full Version)

Ehud Shapiro

REVIEW 1 major objections 1 minor 2 cited by

Implementing Grassroots Logic Programs with Multiagent Transition Systems and AI (Full Version)

T0 review · 1 major / 1 minor · reviewed 2026-08-03 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Grassroots Logic Programs, a nondeterministic concurrent logic language, can be implemented by deterministic FIFO agents exchanging asynchronous messages, preserving outcomes.

desk verdict Serious and mostly coherent formal development, but the headline correctness theorems are not established as stated because the implementation mappings are not well-defined functions on the configuration spaces. read the letter →

arxiv 2602.06934 v6 pith:GMKFKYRJ submitted 2026-02-06 cs.PL cs.AIcs.DCcs.LOcs.MA

Ehud Shapiro This is my paper

classification cs.PLcs.AIcs.DCcs.LOcs.MA

keywords concurrentlogicprogramminggrassrootsplatformsoperationalsemanticsmultiagenttransitionsystemsdeterministicimplementationsingle-occurrenceinvariantasynchronousmessagepassinggloballinks

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to show that Grassroots Logic Programs (GLP), a concurrent logic language with single-reader/single-writer variables and nondeterministic execution, can be implemented by deterministic machines without changing what programs compute. It derives dGLP, a deterministic single-agent semantics, and madGLP, a multiagent semantics in which agents communicate only by asynchronous message passing, and proves both correct against the original nondeterministic semantics. The load-bearing idea is that because every variable is written and read at most once, substitutions produced by different reductions act on disjoint variables and therefore commute; order of execution does not affect the final outcome. If correct, this would mean smartphone, serverless grassroots platforms can run GLP-style logic programs on plain FIFO agents talking over a network.

What carries the argument

The single-occurrence (SO) invariant — each writer and each reader occurs at most once — yields disjoint substitution commutativity: substitutions produced by reductions on different goals assign disjoint variables and therefore commute with each other. Persistence (Lemma 3.17) says an enabled transition stays enabled until an equivalent one is taken. Together these let a deterministic FIFO scheduler reproduce the outcomes of arbitrary nondeterministic runs. In madGLP, the 'global link' implements a shared variable pair across agents via two purely local variable pairs, a global_send goal at one end, and a global writers table entry at the other; the index-0 serializer reduces cold-call conn

What would settle it

Exhibit two different dGLP computations that end in the same triple (active queue, suspended goals, failed goals) but accumulate different reader substitutions; the implementation mapping sigma would then send one configuration to two different GLP states, violating Definition 2.2 and falsifying Theorem 3.34. For madGLP, the analogous test is two identical local states in which different messages are in flight to an agent, mapping to different maGLP states.

Watch

Extended reading notes

Core claim

In the paper's own terms, the contribution is a pair of implementation-ready semantics: dGLP, a deterministic counterpart of the concurrent semantics, and madGLP, a multiagent counterpart where agents communicate only by asynchronous message passing. For each, the paper defines an implementation mapping back to the abstract semantics and proves liveness and completeness, hence correctness (Theorems 3.34 and 5.10). The central structural idea is that GLP's single-occurrence invariant makes substitutions from different reductions commute, so the order in which a deterministic scheduler or an asynchronous network delivers events cannot change the final outcome; this is what lets shared variable

Load-bearing premise

The proofs assume the implementation mapping from each dGLP or madGLP configuration to the abstract state is a well-defined function; but the paper reconstructs reader substitutions from computation history or from messages in channels that the stated configurations do not include, so if those are absent the mapping is not a function and the correctness theorems are not established as stated.

Editorial extensions

If this is right

A nondeterministic, relational logic language can be executed on deterministic hardware without changing program outcomes; goal scheduling order becomes an implementation detail.
Variables shared between agents require no consensus protocol: a writer-to-reader assignment is delivered exactly once via message passing, and the receiver's table routes it to the right local writer.
All multiagent transactions in madGLP are unary; binary atomic handshakes are decomposed into sequences of Reduce, Send, and Receive, which is what lets the protocol be grassroots.
Both madGLP and maGLP satisfy the grassroots property: disjoint groups of agents can run correctly on their own, and any interleaving of their runs is correct, while a cold-call can create genuinely new cross-group behavior.
The formal semantics can serve as a specification for AI-assisted code generation in a math-to-informal-spec-to-code discipline, with implementation feedback correcting the mathematics.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

If the correctness results hold, the global-link decomposition is likely portable to other single-assignment, linear-logic-like languages, giving a general recipe for distributed implementations of logic calculi.
The index-0 serializer pattern, a many-to-one message merge, could be lifted to a user-level language feature, letting programs create additional serialized input channels.
Outcome equivalence is a weaker contract than step-by-step equivalence; users of such an implementation must accept that intermediate interleavings differ even though final results agree.
A formal cleanup would extend dGLP and madGLP configurations with computation history or channel state so the claimed implementation mappings are literally functions on configurations as the paper's Definition 2.2 requires.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The genuinely new thing here is madGLP's global-link encoding: replacing a shared variable pair across agents with two local pairs plus a global_send goal, and showing that the single-occurrence invariant makes the needed substitutions commute. That is a real idea, and the detailed traces in the appendix make the mechanism much easier to follow than the definitions alone would. The single-agent dGLP semantics is also a reasonable deterministic queue-based presentation, with explicit suspension and failure, and the proof strategy based on disjoint substitution commutativity plus persistence is plausible.

The soft spots are formal, and they hit the paper's main claims. Definition 3.28 defines the dGLP implementation mapping from the reduction history, not from the final configuration. A dGLP configuration does not retain consumed reader assignments, so two runs can plausibly reach the same (Q,S,F) with different accumulated substitutions. That is not a function C' -> C as Definition 2.2 requires, so Theorem 3.34 is not proved as stated. The multiagent case is worse: Definition 6.31 reconstructs the maGLP readers substitution only from in-flight messages. Once a message has been received and applied, the assignment is nowhere in madGLP's local state, but maGLP states accumulate assignments monotonically with sent flags. The appendix's Stage 1b is a concrete counterexample: no messages are pending, but the corresponding maGLP state still must contain the delivered assignment. Adding channel state would not fix this; you need history or an explicit accumulated-substitution component in madGLP configurations, or you need to weaken the correctness notion to something like observational equivalence.

I also think the grassroots interactivity proofs are overgeneralized. They assert existence of a cross-group cold-call in a combined run, but the transition system does not show how an agent obtains another agent's address or how the first message appears on the network output stream. For an arbitrary program with no external input, that run may not exist. The paper leans on the in-preparation [50] for the grassroots framework; the definitions are restated well enough, but the interactivity gap is in this paper's own proofs, not just the citation.

These are repair problems, not evidence the construction is wrong. Adding history variables or switching to a run-based implementation relation would likely fix them, and the global-link idea will survive. The paper deserves a serious referee, but only after the mapping definitions are revised. I'd send it to review with that request rather than desk-reject, and I'd tell the authors to be explicit about environment inputs in the grassroots argument.

Referee Report

1 major / 1 minor

Summary. The paper derives two deterministic operational semantics from the nondeterministic concurrent/multiagent logic language GLP: dGLP (single-agent, FIFO-scheduled with suspension/failure) and madGLP (multiagent, asynchronous message passing over global links). It claims dGLP correctly implements GLP (Theorem 3.34) and madGLP correctly implements maGLP (Theorem 5.10), using an implementation-mapping framework from Section 2, and that both madGLP and maGLP are grassroots (Theorems 8.9 and 8.10). The paper also reports an AI-assisted development pipeline from the mathematical specifications to Dart implementations.

Significance. If the correctness results were fully established, the paper would make a useful contribution: it would show that a nondeterministic concurrent logic calculus with paired reader/writer variables can be implemented by deterministic FIFO agents and asynchronous message passing without changing outcomes. The central proof ideas — disjoint substitution commutativity from the single-occurrence invariant, and persistence — are plausible and elegant, and the detailed traces in Appendix A are helpful in understanding the intended design. The paper ships no machine-checked proofs, and several key definitions are borrowed from unpublished or in-preparation work ([42], [46], [50]), so the self-containedness of the correctness argument is limited. The main issue is that the implementation mappings used in the central theorems are not well-defined functions on the stated configuration spaces.

major comments (1)

[Overall] Given that the implementation mappings are not well-defined as stated and the Send/Receive transitions reference absent channel state, the central correctness theorems need revision. I recommend major revision.

minor comments (1)

[General] There are a few typos and formatting artifacts (e.g., 'envel⌢pe/orcid' in the author line, the repeated reference to 'Definition 4.2' in the proof of Lemma 3.30 where it should be 'Definition 3.11'). These are minor and do not affect the technical content.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: dGLP/madGLP correctness and grassroots results are proved from definitions via refinement mappings; the formal gaps identified are soundness/well-definedness issues, not circular reductions.

full rationale

The paper's derivation chain is not circular in the sense of assuming its own conclusions. dGLP (Definition 3.25) is an independent FIFO/suspend/fail transition system, and Theorem 3.34 is proved from Lemmas 3.29–3.33 using only the stated GLP and dGLP definitions. The implementation mapping σ (Definition 3.28) is a refinement mapping; its reliance on computation history is a well-definedness defect relative to Definition 2.2, but it is not a case of equating the theorem with an input. Likewise, madGLP correctness (Theorem 5.10) is argued through the constructed mapping π (Definitions 6.30–6.32) and correspondence lemmas (6.28–6.38). The reconstruction of σ_p only from in-flight messages in Definition 6.31 is too weak—delivered assignments vanish once messages leave the channel (cf. Appendix A.1 Stage 1b, where no messages are pending after receipt but the maGLP state retains the assignment)—but this is a soundness/proof gap, not a circular equation. The grassroots theorems (8.9, 8.10) are proved from the paper's own Definition 8.4 and Theorem 8.8; the citation to [50] supplies terminology/definitional context that the paper reproduces, so the self-citation is not load-bearing. No parameters are fitted, no prediction is defined as its input, and no central claim reduces to a prior self-citation by construction. The numerous self-citations, including the placeholder 'in preparation' reference [50], are a missing-support/reference concern rather than circularity.

Assumptions & free parameters 0 free parameters · 5 assumptions · 3 invented entities

The paper's central claims rest on the author's prior GLP semantics, an unpublished grassroots/monotonic-transactions framework, and new madGLP runtime machinery (global_send, global writers table, index-0 serializer). No free parameters are fitted; the invented entities are implementation mechanisms rather than empirical postulates, but none is backed by released code or tests.

assumptions (5)

domain assumption GLP syntax and abstract semantics (single-occurrence invariant, SRSW, term-matching writer mgu, GLP transition system) are accepted as given from [46].
All correctness proofs are relative to this language definition; it is the author's prior preprint, not re-derived here.
domain assumption Liveness/fairness framework and Definition 2.2–2.3 of implementation correctness are assumed from [42].
Correctness is defined as live-plus-complete using fair runs, but the general liveness notion is not recalled in the paper.
domain assumption The grassroots property is defined as oblivious + interactive per [50], an in-preparation manuscript by the same author.
Theorems 8.9 and 8.10 depend on this definition; it cannot be independently checked from this paper.
domain assumption madGLP assumes fair message delivery on reliable communication channels.
Used implicitly in Send/Receive correctness and liveness (Section 5.2, Lemma 6.36).
ad hoc to paper Initial local states use anonymous variables '_' that are syntactically identical but denote fresh unique variables in each agent.
Invented to satisfy Definition 4.3's identical s0 requirement while allowing distinct local variables; not an ordinary GLP term.

invented entities (3)

global_send/3 system predicate
purpose: Monitors a local reader and sends an outbound message once the reader is assigned a value; implements the outgoing half of global links.
A new builtin defined in Definition 6.5; no implementation or test artifacts are shipped, only an informal spec claimed.
global writers table with entries (X,q) and (X,q,i)
purpose: Routes incoming assignment messages to the correct local writer at the receiving agent.
New runtime datastructure (Definitions 6.2–6.3); no code released.
index-0 serializer _r(p,0)
purpose: Allows multiple agents to send cold-call messages to one agent by appending to its network input stream.
New mechanism (Definition 6.7) with no external falsifiable handle beyond the informal specification.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Implementing Grassroots Logic Programs with Multiagent Transition Systems and AI (Full Version)." pith.science (2026). https://pith.science/paper/GMKFKYRJ

@misc{pith2026260206934,
  author       = {Pith},
  title        = {Pith review of: Implementing Grassroots Logic Programs with Multiagent Transition Systems and AI (Full Version)},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GMKFKYRJ}},
  note         = {Machine review of arXiv:2602.06934}
}

abstract

Grassroots Logic Programs (GLP) is a concurrent logic programming language in which logic variables are partitioned into paired readers and writers. An assignment is produced at most once via a writer and consumed at most once via its paired reader, and may contain additional readers and/or writers. This enables the concise expression of rich multidirectional communication modalities. The language was introduced together with concurrent (cGLP) and multiagent (maGLP) operational semantics. Here, we derive from these (\ia)~dGLP, a deterministic counterpart of cGLP, and (\ib)~madGLP, a counterpart of maGLP in which deterministic agents communicate solely by asynchronous message passing, and prove them correct against their abstract counterparts. maGLP shared variable pairs spanning agents can be implemented by two local variable pairs joined by a \emph{global link}, with correctness following from disjoint substitution commutativity (a consequence of GLP's single-occurrence invariant). We further prove that madGLP is grassroots. Both dGLP and madGLP serve as formal specifications for an AI-driven implementation discipline (math $\to$ informal spec $\to$ Dart) employed and described here: from dGLP, AI (Claude) developed a workstation-based GLP implementation in Dart, and from madGLP it is developing a smartphone-based multiagent one.

Review history (2 revisions) →

Discussion (0). Sign in to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

Volition Elicitation: Operational Semantics for People and Their Machines
cs.PL 2026-07 conditional novelty 6.0 of 10

vGLP extends GLP so that program reductions can be guarded by a person's expressed volition, with the UI derived from the semantics and correctness proofs for soundness, liveness, and compilation.
GLP: A Grassroots, Multiagent, Concurrent, Logic Programming Language for AI
cs.PL 2026-07 conditional novelty 5.0 of 10

GLP is a single-assignment concurrent logic language whose multiagent semantics is claimed to guarantee that any program using cold-calls yields a grassroots platform, with the proof deferred to the full paper.

Reference graph

Works this paper leans on

24 extracted references · 6 linked inside Pith · cited by 2 Pith papers

[1]

The existence of refinement mappings.Theoretical Computer Science, 82(2):253–284, 1991.doi:10.1016/0304-3975(91)90224-P

1 Martín Abadi and Leslie Lamport. The existence of refinement mappings.Theoretical Computer Science, 82(2):253–284, 1991.doi:10.1016/0304-3975(91)90224-P. 2 Paulo Sérgio Almeida and Ehud Shapiro. The blocklace: A byzantine-repelling and universal conflict-free replicated data type.arXiv preprint arXiv:2402.08068,

arXiv 1991
[5]

15Google

doi: 10.1016/0304-3975(87)90045-4. 15Google. Dart programming language.https://dart.dev,

doi:10.1016/0304-3975(87)90045-4
[16]

45 Ehud Shapiro

doi:10.1145/3599696.3612898. 45 Ehud Shapiro. Grassroots currencies: Foundations for grassroots digital economies.arXiv preprint arXiv:2202.05619,

arXiv
[17]

Glp: A grassroots, multiagent, concurrent, logic programming language.arXiv preprint arXiv:2510.15747,

46 Ehud Shapiro. Glp: A grassroots, multiagent, concurrent, logic programming language.arXiv preprint arXiv:2510.15747,

arXiv
[19]

51 Ehud Shapiro and Nimrod Talmon

In preparation. 51 Ehud Shapiro and Nimrod Talmon. Grassroots federation: Fair governance of large-scale, decentralized, sovereign digital communities.arXiv preprint arXiv:2505.02208,

arXiv
[20]

Conflict-free replicated data types

52 Marc Shapiro, Nuno Preguiça, Carlos Baquero, and Marek Zawirski. Conflict-free replicated data types. InStabilization, Safety, and Security of Distributed Systems: 13th International Symposium, SSS 2011, Grenoble, France, October 10-12,

2011
[22]

Resource-passing concurrent programming.Proceedings of TACS 2001, pages 95–126,

56 Kazunori Ueda. Resource-passing concurrent programming.Proceedings of TACS 2001, pages 95–126,

2001
[23]

A taste of linear logic.Mathematical Structures in Computer Science, 3(4):367– 392, 1993.doi:10.1017/S0960129500000268

58 Philip Wadler. A taste of linear logic.Mathematical Structures in Computer Science, 3(4):367– 392, 1993.doi:10.1017/S0960129500000268. Ehud Shapiro 35 59 Philip Wadler. Propositions as sessions. InProceedings of the 17th ACM SIGPLAN In- ternational Conference on Functional Programming (ICFP), pages 273–286. ACM,

doi:10.1017/s0960129500000268 1993

Show all 24 references

[24]

60 Philip Wadler

doi:10.1145/2364527.2364568. 60 Philip Wadler. Propositions as sessions.Journal of Functional Programming, 24(2–3):384–418, 2014.doi:10.1017/S095679681400001X. A Detailed madGLP Example Traces This appendix provides complete formal traces for three madGLP examples: the client-...

2014
[1985]

Grassroots flash: A payment system for grassroots cryptocurrencies.arXiv preprint arXiv:2309.13191,

23 Andrew Lewis-Pye, Oded Naor, and Ehud Shapiro. Grassroots flash: A payment system for grassroots cryptocurrencies.arXiv preprint arXiv:2309.13191,

arXiv
[1987]

Statically contextualizing large language models with typed holes

6 Andrew Blinn, Xiang Li, June Hyung Kim, and Cyrus Omar. Statically contextualizing large language models with typed holes. InProceedings of the ACM on Programming Languages (OOPSLA), volume 8, pages 1–29. ACM, 2024.doi:10.1145/3689746. 7 Luís Caires and Frank Pfenning. Sessi...

2024 doi
[1989]

Multiagent transition systems: Protocol-stack mathematics for distributed computing.arXiv preprint arXiv:2112.13650,

42 Ehud Shapiro. Multiagent transition systems: Protocol-stack mathematics for distributed computing.arXiv preprint arXiv:2112.13650,

arXiv
[1991]

40 Ehud Shapiro

doi:10.1145/99583.99627. 40 Ehud Shapiro. A subset of concurrent prolog and its interpreter.ICOT Technical Report, TR-003,
[1994]

Lynch and Mark R

28 Nancy A. Lynch and Mark R. Tuttle. Hierarchical correctness proofs for distributed algorithms. InProceedings of the 6th Annual ACM Symposium on Principles of Distributed Computing (PODC), pages 137–151. ACM, 1987.doi:10.1145/41840.41852. 29 Nancy A. Lynch and Mark R. Tuttle...

1987
[1995]

31 NancyA.LynchandFritsW.Vaandrager

doi:10.1006/inco.1995.1134. 31 NancyA.LynchandFritsW.Vaandrager. Forwardandbackwardsimulations: II.Timing-based systems.Information and Computation, 128(1):1–25, 1996.doi:10.1006/inco.1996.0060. 32 Nicholas D. Matsakis and Felix S. Klock II. The Rust language. InProceedings of...

1995
[2010]

4 Henry G

doi:10.1145/ 1810891.1810910. 4 Henry G. Baker and Carl Hewitt. The incremental garbage collection of processes. In Proceedings of the 1977 Symposium on Artificial Intelligence and Programming Languages, pages 55–59. ACM, 1977.doi:10.1145/800228.806932. 5 Philip A. Bernstein, ...

1977
[2011]

Higher-order processes, functions, and sessions: A monadic integration

53 Bernardo Toninho, Luís Caires, and Frank Pfenning. Higher-order processes, functions, and sessions: A monadic integration. InProceedings of the 22nd European Symposium on Programming (ESOP), volume 7792 ofLecture Notes in Computer Science, pages 350–369. Springer, 2013.doi:...

2013 doi
[2012]

37 Klaas Pruiksma and Frank Pfenning

Available athttps://go.dev/talks/2012/splash.article. 37 Klaas Pruiksma and Frank Pfenning. Back to futures.Journal of Functional Programming, 32:e4, 2022.doi:10.1017/S0956796822000016. 34 Implementing Grassroots Logic Programs 38 Davide Sangiorgi. On the bisimulation proof me...

2012 doi
[2015]

10 Ezra Cooper, Sam Lindley, Philip Wadler, and Jeremy Yallop

doi:10.1145/2824815.2824816. 10 Ezra Cooper, Sam Lindley, Philip Wadler, and Jeremy Yallop. Links: Web programming without tiers. InProceedings of the 5th International Symposium on Formal Methods for Components and Objects (FMCO), volume 4709 ofLecture Notes in Computer Scien...

2006
[2021]

Grassroots distributed systems: Concept, examples, implementation and applications (brief announcement)

43 Ehud Shapiro. Grassroots distributed systems: Concept, examples, implementation and applications (brief announcement). In37th International Symposium on Distributed Computing (DISC 2023). (Extended version: arXiv:2301.04391), pages 47:1, 47:7, Italy,

2023
[2023]

Practical type inference based on success typings

24 Tobias Lindahl and Konstantinos Sagonas. Practical type inference based on success typings. InProceedings of the 8th ACM SIGPLAN International Conference on Principles and Practice of Declarative Programming (PPDP), pages 167–178. ACM, 2006.doi:10.1145/1140335. 1140356. 25 ...

2006 doi
[2024]

Herlihy and Jeannette M

17 Maurice P. Herlihy and Jeannette M. Wing. Linearizability: A correctness condition for concurrent objects.ACM Transactions on Programming Languages and Systems, 12(3):463– 492, 1990.doi:10.1145/78969.78972. Ehud Shapiro 33 18 Joshua S. Hodas and Dale Miller. Logic programmi...

1990
[2025]

36 Rob Pike

Available athttps://arxiv.org/ abs/2504.09246. 36 Rob Pike. Go at Google: Language design in the service of software engineering. Keynote at SPLASH 2012,

2012 arXiv
[2026]

doi: 10.1145/3772290.3772309

arXiv preprint arXiv:2502.11299. doi: 10.1145/3772290.3772309. 49Ehud Shapiro. Types for grassroots logic programs.arXiv preprint arXiv:2601.17957,

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.

[1] [1]

The existence of refinement mappings.Theoretical Computer Science, 82(2):253–284, 1991.doi:10.1016/0304-3975(91)90224-P

1 Martín Abadi and Leslie Lamport. The existence of refinement mappings.Theoretical Computer Science, 82(2):253–284, 1991.doi:10.1016/0304-3975(91)90224-P. 2 Paulo Sérgio Almeida and Ehud Shapiro. The blocklace: A byzantine-repelling and universal conflict-free replicated data type.arXiv preprint arXiv:2402.08068,

arXiv 1991

[2] [5]

15Google

doi: 10.1016/0304-3975(87)90045-4. 15Google. Dart programming language.https://dart.dev,

doi:10.1016/0304-3975(87)90045-4

[3] [16]

45 Ehud Shapiro

doi:10.1145/3599696.3612898. 45 Ehud Shapiro. Grassroots currencies: Foundations for grassroots digital economies.arXiv preprint arXiv:2202.05619,

arXiv

[4] [17]

Glp: A grassroots, multiagent, concurrent, logic programming language.arXiv preprint arXiv:2510.15747,

46 Ehud Shapiro. Glp: A grassroots, multiagent, concurrent, logic programming language.arXiv preprint arXiv:2510.15747,

arXiv

[5] [19]

51 Ehud Shapiro and Nimrod Talmon

In preparation. 51 Ehud Shapiro and Nimrod Talmon. Grassroots federation: Fair governance of large-scale, decentralized, sovereign digital communities.arXiv preprint arXiv:2505.02208,

arXiv

[6] [20]

Conflict-free replicated data types

52 Marc Shapiro, Nuno Preguiça, Carlos Baquero, and Marek Zawirski. Conflict-free replicated data types. InStabilization, Safety, and Security of Distributed Systems: 13th International Symposium, SSS 2011, Grenoble, France, October 10-12,

2011

[7] [22]

Resource-passing concurrent programming.Proceedings of TACS 2001, pages 95–126,

56 Kazunori Ueda. Resource-passing concurrent programming.Proceedings of TACS 2001, pages 95–126,

2001

[8] [23]

A taste of linear logic.Mathematical Structures in Computer Science, 3(4):367– 392, 1993.doi:10.1017/S0960129500000268

58 Philip Wadler. A taste of linear logic.Mathematical Structures in Computer Science, 3(4):367– 392, 1993.doi:10.1017/S0960129500000268. Ehud Shapiro 35 59 Philip Wadler. Propositions as sessions. InProceedings of the 17th ACM SIGPLAN In- ternational Conference on Functional Programming (ICFP), pages 273–286. ACM,

doi:10.1017/s0960129500000268 1993