Functional graph reduction: implementation evidence
← Return to the implementation matrix
A graph-reduction machine evaluates a functional program by rewriting its expression graph. The credited hardware line is narrower than functional programming generally.
| Stage | Mark | Summary |
|---|---|---|
| Reference | ● | Executable reducer designs and toolchains |
| Physical | ● | FPGA processor core |
| Integrated | ● | Reduction machinery assembled as a programmable core |
| Scaled | — | No many-core reducer demonstrated |
| Access | ◐ | Research artifacts and FPGA-oriented designs |
| Operational | — | No recurring deployment located |
Reference — demonstrated
- Claim
- The matrix credits functional graph reduction at the Reference stage: executable reducer designs and toolchains.
- Evidence
- The Heron and HAFLANG work specifies executable reduction machinery for pure, non-strict functional languages. The design makes allocation, graph traversal, and reduction explicit enough to compile and test.
- Criticism
- Executable semantics do not establish purpose-built hardware, integration, scale, external access, recurring use, or comparative advantage.
- Sources
- Heron: Modern Hardware Graph Reduction; HAFLANG
Physical — demonstrated
- Claim
- The matrix credits functional graph reduction at the Physical stage: FPGA processor core.
- Evidence
- Heron was implemented as an FPGA processor core. That crosses the physical threshold: graph-reduction operations are realized by purpose-built logic rather than only interpreted by a CPU.
- Criticism
- A physical realization does not by itself establish system integration, efficient scaling, external access, recurring use, or comparative advantage.
- Sources
- Heron: Modern Hardware Graph Reduction
Integrated — demonstrated
- Claim
- The matrix credits functional graph reduction at the Integrated stage: reduction machinery assembled as a programmable core.
- Evidence
- The FPGA core brings reduction control, graph storage operations, and execution together as a machine. The mark is for a coherent research processor, not for a complete stand-alone computer with mature storage and I/O.
- Criticism
- A coherent system does not by itself establish efficient scaling, external access, recurring use, or comparative advantage.
- Sources
- Heron: Modern Hardware Graph Reduction
Access — limited
- Claim
- The matrix credits functional graph reduction only partially at the Access stage: research artifacts and FPGA-oriented designs.
- Evidence
- The architecture and research line are published, and HAFLANG continues work on hardware acceleration. Access remains limited because there is no maintained public service or generally available graph-reduction appliance.
- Criticism
- The mark is limited on the current public record: research artifacts and FPGA-oriented designs. Access does not establish broad availability, recurring use, workload generality, or comparative advantage.
- Sources
- Heron: Modern Hardware Graph Reduction; HAFLANG
Stages not credited
No public evidence establishes a many-core reducer whose allocation, communication, and garbage-collection costs remain favorable at scale. No recurring operational deployment was located.