Product Spark vs Forge Spark

Product Spark is a shippable product slice (PoC/MVP/phase); Forge Spark is a single executable task (~1–4h) in Charge — maps to WBS Task. See NAMING-REFERENCE.md.

The collision

Plans size a "spark" afternoon task as a quarter roadmap, or label learning spikes as Sparks — duplicate hierarchies and broken Assay/WBS traceability.

How to choose

1) Shippable increment with Assay / milestone M1? → Product Spark (release slice). 2) Today's executable task in Charge? → Forge Spark (phase-prefixed task id). 3) Time-boxed learning only? → discipline spike (spike_discipline) — not either Spark type. 4) Colloquial "product spike"? → discipline spike with PM lens — not Product Spark.

Using several at once

Product Spark rolls up Ingots and Forge Sparks via WBS; one ID spine when using M1E1S1T1 pattern. forge repo abbr is unrelated to Forge Spark task noun.

Terms covered

Examples from chat / plan.md

"M1 MVP Spark" in roadmap → Product Spark named M1; tasks discover:oauth-config and specify:api-contract are Forge Sparks.

Exploration spike for CDP attach feasibility → spike_discipline in forge-logs — not a Forge Spark committed to Charge.


Bridge page product-spark-vs-forge-spark — read when multiple abbreviations appear in one sentence.