Solar on every roof, BOI solar correction, open source section

Solar (user, 17 Oct 2026: "use as much roof space as possible")
- PV on every roof and canopy: ~640 kWp on the Payap scenario plot
  (factory roof ~523, canopies ~54, flat office roof ~65 kWp around the
  AC units); ~300 kWp on the Mueang Chi hall. Calculator: pvKWp 640,
  new pvSelf 85 % (zero export, hourly sketch), PV cost per kWp times the
  array, office-roof PV in the iso view, roofs drawn full.
- 12.8M to 16.6M THB at 20 to 26 THB/W; CSV baseline 34.37M to 56.34M;
  launch plant ~168M / 231M / 294M THB with contingency; Plan B ~216M
  (label ~215M), ~39M above Plan A; Plan A unchanged (it leases the
  array), its solar lever ~17.5M. PV payback ~4.2 to 5.9 years.
- Permits (rooftop solar no factory licence since 28 Dec 2024, ERC
  exemption under 1,000 kVA, PEA zero export) and PPA discounts sourced
  in sustainability.md.

BOI correction
- Announcement 3/2568 (in force 1 July 2025) ended BOI promotion for
  solar installations; only the battery may qualify (12M THB per MW cap).
  Removed the "BOI solar measure" from boi.md, total-investment,
  revenue-model, sustainability, site-candidates and both plan versions;
  Plan B's case for owning the array is the PPA bill it avoids
  (~0.07M USD a year).

Open source (user, 17 Oct 2026)
- digital-factory.md "Open source throughout, and working with other
  plants": everything we run and build is open source on NixOS, our code
  published (licence to decide), vendor firmware the stated exception;
  later the stack and factory-automation services for other pressing
  plants, not in the revenue.
- Business plan section 5.6 plus an executive-summary bullet, reading
  guide and not-modelled mentions; draft 0.3 rebuilt (A 60, B 61 pages),
  renders re-made with the roofs full.

docs/roadmap.md

@@ -11,7 +11,7 @@ These three run in parallel and gate everything downstream:
 1. **Market validation**: test the blended pressing price (3.76 to 5.27
    USD at the planned 60 % specials; 3.50 to 4.60 at a standard-heavy mix)
    and regional demand (~0.74M records a year in both Plan A and Plan B, and
-   when the second Duplex follows from revenue) in [total investment](finance/total-investment.md#two-plans-a-at-180m-thb-and-b-at-210m-thb-october-2026),
+   when the second Duplex follows from revenue) in [total investment](finance/total-investment.md#two-plans-a-at-180m-thb-and-b-at-215m-thb-october-2026),
    with Asian labels and distributors, starting at
    Making Vinyl in Haarlem (Alexander and Simona). This is the biggest
    assumption in the [revenue model](finance/revenue-model.md); the
@@ -173,9 +173,9 @@ Nothing in Phases 2 to 5 starts before its gate clears: no machinery order
 before the BOI certificate, no big-block spending before the market
 validates the price, no coppice fuel planning before the land title check.
 Since October 2026 the plant launches with everything in one capital
-round (~160M to 283M THB with contingency; for the investor Plan A at
-~180M and Plan B at ~210M THB without it, in
-[total investment](finance/total-investment.md#two-plans-a-at-180m-thb-and-b-at-210m-thb-october-2026));
+round (~168M to 294M THB with contingency; for the investor Plan A at
+~180M and Plan B at ~215M THB without it, in
+[total investment](finance/total-investment.md#two-plans-a-at-180m-thb-and-b-at-215m-thb-october-2026));
 the gates still come before the round is spent, and the growth options
 (more Duplex systems first, then Classics, in a press hall laid out for
 five Duplex systems and four Classics) are funded from revenue.