====== Concepts ======

  * [[brainstorming_initial]]

While the market is already oversaturated with cheap, unreliable IoT irrigation products and AI irrigation fluff, our goal is to deliver not just some cheap hardware, but a well developed, robust and reliable and maintainable solution. 

Our System needs to be able to deal well with:

  * bad power
  * no internet
  * heat & dust
  * unqualified field workers

So this means, our Architecture needs to be:

  * offline-first
  * hardware-enforced safety
  * deterministic control
  * supervisory only networking
  * explicit decision precedence
  * audit-ready by design

And our Architecture also needs to be able to:

  * work after 3 years (most cheap smart farming stuff does not)
  * survive a dead battery
  * survive a broken gateway
  * survive an technician swapping wires
  * survive being offline for months

===== Business Model =====

==== We do NOT want to build this as a product ====
What we do NOT want to do is to build this as a "product":

  * hardware margins are very low
  * support is much effort
  * install quality dominates outcomes
  * we well get undercut by junk vendors immediately
  * farmers don't pay for architecture elegance

If we try to sell devices, we lose.

==== What we should try to aim for ====

=== Reference Architecture ===
We can publish this as:

  * a donor-compliant reference design
  * a minimum safety & control standard
  * an open or semi-open architecture

Then:

  * integrators build on it
  * governments reference it
  * bad products get filtered out

This would make us the //author//, not the //vendor//.

=== Compliance Layer ===
We do not focus on building hardware (except some reference implementations), but we focus on building:

  * control rules
  * safety invariants
  * validation tests
  * audit schemas

Vendors must prove: //"Our system conforms to this architecture"//.

This is how we avoid the race to the bottom of "we can build this box even cheaper than the others!".


What we sell is:

  * design review
  * system validation
  * commissioning rules
  * Monitoring & Evaluation compatibility