docs: rewrite Phlo Transform blog from first principles

[iamgp]

Summary

Reworks the Phlo Transform blog into a much more complete first-principles series aligned with the v0.1 implementation.

Naming rule

The series now uses Phlo Transform consistently for this product. Phlo is reserved for the wider lakehouse framework and is no longer used as shorthand for Phlo Transform in prose. Literal implementation identifiers such as phlo-transform, phlo.toml, .phlo/, phlo_main, and crate names remain unchanged.

Rewritten

• 01 — why SQL becomes a build-system problem; defines models, relations, materialisation, desired/current state
• 02 — compiler pipeline, parsing, resolution, logical vs physical names, directives and source qualification
• 03 — one workspace graph, DAGs, typing, column lineage, impact, contracts, keys and tests
• 04 — state, semantic versioning, live physical identity, planning, BUILD/SKIP/CACHED, verified cross-environment adoption
• 05 — incremental strategies, watermarks, failure-safe state and cache adoption
• 06 — Iceberg/Nessie environments, WAP, diffs, portable immutable evidence, exact evidence IDs and promotion/cleanup semantics
• 07 — dbt migration as semantic translation with CLEAN/REVIEW/UNSUPPORTED

Added

• 08 — execution engine: DAG scheduling, bounded concurrency, seeds/tests, retries, cancellation, resume and retry-failed
• 09 — daemon/machine API: read vs operation surfaces, idempotency, durable journal, auth and agent usage
• 10 — end-to-end v0.1 lifecycle from blank workspace through candidate cache adoption, audit and promotion

Index

docs/blog/README.md now includes a reading guide, glossary, all 10 posts, and the current v0.1 features reflected across the series.

Docs only; no runtime code changes.