tag

#research

18 pages tagged

Practice › Mathematics › TAOCP › TAOCP Vol 1: Fundamental Algorithms › TAOCP 2.2.3: Linked Allocation ›

TAOCP 2.2.3 Exercise 25

Section 2.2.3: Linked Allocation Exercise 25. [ 47 ] Design as efficient an algorithm as possible for doing a topological sort of very large sets $S$ having considerably more nodes than the computer memory can contain. Assume that the input, output, and temporary working space are done with magnetic tape. [ Possible hint: A conventional sort of the input allows us to assume that all relations for a given node...

Practice › Mathematics › TAOCP › TAOCP Vol 1: Fundamental Algorithms › TAOCP 1.2.1: Mathematical Induction ›

TAOCP 1.2.1 Exercise 14

Section 1.2.1: Mathematical Induction Exercise 14. [ 50 ] (R. W. Floyd.) Prepare a computer program that accepts, as input, programs in some programming language together with optional assertions, and that attempts to fill in the remaining assertions necessary to make a proof that the computer program is valid. (For example, strive to get a program that is able to prove the validity of Algorithm E, given only assertions A1...

Research › Mochi-to-C Transpiler ›

Codegen design

MEP-45 research note 05, Codegen design Author: research pass for MEP-45. Date: 2026-05-22 (GMT+7). This note specifies the shape of the C the transpiler emits and the IR it uses to get there. It is the technical heart of MEP-45 and is intended to be a working blueprint. Lowering of the eight sub-languages is described in note 01 §1–10; this note covers the cross-cutting mechanics. 1. Pipeline overview .mochi source...

Research › Mochi-to-C Transpiler ›

Risks and alternatives

MEP-45 research note 12, Risks and alternatives Author: research pass for MEP-45. Date: 2026-05-22 (GMT+7). This note is the "what could go wrong, and what would we do instead" discussion that PEP-style proposals are expected to carry. 1. Risks 1.1 Semantic divergence Risk: the C target diverges from the VM on edge cases (integer overflow, float NaN handling, string comparison, map iteration ordering, GC observable behaviour). Mitigation: byte-equal differential testing...

Research › Mochi-to-C Transpiler ›

Language surface

MEP-45 research note 01, Mochi language surface Author: research pass for MEP-45 (Mochi → C transpiler). Date: 2026-05-22 (GMT+7). Sources: docs/features/*.md , docs/index.md , docs/common-language-errors.md , mcp/cheatsheet.mochi , ROADMAP.md , examples/v0.2 – v0.7 , normative security specs docs/security/threat-model.md and docs/security/memory-safety.md . This note records the user-visible language surface that the C target must faithfully reproduce. It is deliberately written from the spec downward and ignores the existing Go runtime, the...

Research › Mochi-to-C Transpiler ›

Dataset pipeline lowering

MEP-45 research note 08, Dataset pipeline lowering Author: research pass for MEP-45. Date: 2026-05-22 (GMT+7). Mochi's headline feature is the query DSL on collections. The same syntax operates on lists, maps with insertion-ordered iteration, and streams of records loaded from JSON, YAML, CSV, JSONL, or a load adapter. This note covers the lowering strategy. 1. Language surface recap From the language docs and examples/load_with.mochi : let people = load "people.json"...

Research › Mochi-to-C Transpiler ›

Prior-art transpilers

MEP-45 research note 03, Prior art: language → C compilation (2024–2026) Author: research pass for MEP-45. Date: 2026-05-22 (GMT+7). Method: structured web research by a delegated agent; verbatim report preserved below with light section-anchor edits. The report is the canonical survey for the MEP body's "Rationale" and "Prior Art" sections. References at the foot are the authoritative source list. Survey: State of the Art in Compiling High-Level Languages to C...

Research › Mochi-to-C Transpiler ›

Runtime building blocks

MEP-45 research note 04, Runtime building blocks Author: research pass for MEP-45. Date: 2026-05-22 (GMT+7). This note inventories the third-party and home-grown components the C runtime can stand on. It does not prescribe which goes into v1 (that is the MEP body); it captures what is actually available in 2026, with licence, portability, and integration cost recorded so that the MEP can defend its choices. 1. Memory management 1.1 Conservative...

Research › Mochi-to-C Transpiler ›

Streams and agents

MEP-45 research note 09, Streams and agents Author: research pass for MEP-45. Date: 2026-05-22 (GMT+7). Mochi has first-class stream<T> values, stream definitions that declare event types, on handlers that subscribe to events, and agent records with intent methods that wrap a mailbox-style actor model. This note covers their lowering. 1. Surface recap From examples/v0.4/stream.mochi and examples/v0.5/agent.mochi and the language docs: stream prices { symbol: string price: float at: time }...

Research › Mochi-to-C Transpiler ›

C target and portability

MEP-45 research note 07, The C target and portability Author: research pass for MEP-45. Date: 2026-05-22 (GMT+7). This note covers the output side of the transpiler. What flavour of C we emit, what compilers must accept it, which architectures and OSes get binary artefacts, what hardening is on by default, and what style rules the generated code follows so that a reviewer (or a tool like clang-tidy ) sees consistent...

Research › Mochi-to-C Transpiler ›

Design philosophy

MEP-45 research note 02, Design philosophy Author: research pass for MEP-45. Date: 2026-05-22 (GMT+7). This note records the principles the C target should be designed against, written before any code is touched, so that the implementation can be audited against them later. The philosophy is distilled from three places: the language surface (note 01), the Mochi project's own existing priorities ( README.md "small, statically typed, … zero-dependency single binary, …...

Research ›

Mochi-to-C Transpiler

Background research for Mochi MEP-45 : the deep-dive specification for the C-as-target AOT half of MEP-42 . The transpiler takes compiler3 IR, lowers it to ISO C23 plus a thin libmochi.a runtime, and ships a statically-linked single-file native binary for every tier-1 triple (x86_64-linux-{gnu,musl}, aarch64-linux-{gnu,musl}, aarch64-darwin, x86_64-darwin, x86_64-windows-msvc, x86_64-windows-gnu, wasm32-wasi). The master correctness gate is byte-equal stdout from the produced binary versus vm3 on the entire fixture corpus. Each file...

Research › Mochi-to-C Transpiler ›

Testing and CI gates

MEP-45 research note 11, Testing and CI gates Author: research pass for MEP-45. Date: 2026-05-22 (GMT+7). The single most important property a transpiler can have is output correctness . This note describes how MEP-45 proves it. 1. Differential testing against the VM Every fixture under tests/vm/valid/ (and the examples/v0.3/ ... examples/v0.5/ corpora) has a recorded expected stdout produced by the VM. The transpiler's output must: Compile under tier-1 toolchains. Run...

Research › Mochi-to-C Transpiler ›

Build system

MEP-45 research note 10, Build system Author: research pass for MEP-45. Date: 2026-05-22 (GMT+7). This note covers the build driver. The driver is the thing that takes a Mochi source tree and turns it into a binary. The transpiler is one stage of the driver. 1. Command surface mochi build [PATH] # build the current module mochi build --target=x86_64-linux-musl PATH # cross-compile mochi build --target=wasm32-wasi PATH # WASM/WASI mochi build...

Research › Mochi-to-C Transpiler ›

Type-system lowering

MEP-45 research note 06, Type-system lowering details Author: research pass for MEP-45. Date: 2026-05-22 (GMT+7). Companion to note 05. Focused on the type-system corners and how they become C types and C invariants. This note assumes the codegen scaffold in note 05 and elaborates one piece at a time. 1. Generics and monomorphisation Mochi's surface generics are limited to type parameters on collections and on user record/sum types (no higher-kinded...

Research ›

Native Code Emission

Background research for Mochi MEP-42 : the native-codegen positioning that pairs a copy-and-patch JIT (Xu+Kjolstad PLDI 2021, validated by CPython 3.13 in October 2024) for mochi run hot loops with a C-as-target AOT pipeline (Nim / V / Vala lineage) for mochi build distributables. Each subsection drills into one thread of the 2024-2026 native-codegen landscape. Every file has a §1 Provenance with canonical URLs, a §2 Mechanism, a §3 status...

Research

Long-form research notes that back specific design documents. Each subsection collects the literature, systems studies, and provenance for one decision-making artifact. Subjects Memory Management — the substrate for Mochi MEP-41 : capability machines, generational references, ownership systems, runtime techniques, formal verification, and 2026 industry policy. Native Code Emission — the substrate for Mochi MEP-42 : copy-and-patch JIT, C-as-target AOT, target ISAs and ABIs, object formats, linkers, runtime, debug info, and...

Research ›

Memory Management

Background research for Mochi MEP-41 : the memory-safety positioning that frames Mochi's vm3 runtime as a Vale generational reference machine, an MSWasm capability ABI, and a kalloc_type / xzone typed allocator, all under one roof. Each subsection drills into one thread of the 2023-2026 memory-safety landscape. Every file has a §1 Provenance with canonical URLs, a §2 Mechanism, a §3 status as of May 2026, an engineering-cost section, a Mochi...