Principles are only useful if they can tell us we are wrong. Each one below is written so that it can be violated, and so that we would be able to notice.
1. Measure first, claim second
We do not assert an impact we have not measured. The sequence is: measure the starting point → establish the need → state the expected result → implement → measure the actual result. This is specified in the Evidence protocol.
Violation looks like: a proposal whose benefit is expressed only as a percentage from a manufacturer’s brochure.
2. Label the strength of every claim
Everything we state carries an evidence tier — from measured on a site we instrumented down to heard from a neighbour. Both are legitimate inputs. Presenting the second as the first is not. Tiers are defined in the Source registry.
Violation looks like: a number in a client document with no traceable source.
3. Publish what does not work
Negative and disappointing results are published alongside successes. They are more valuable to a reader deciding where to spend money, and they are the only credible signal that our positive results are real.
Violation looks like: a case study library in which every intervention succeeded.
4. Depth before breadth
We would rather know one domain to the bottom than seven superficially. Breadth comes from repeating depth, not from skipping it.
Violation looks like: six domain briefs all at 20% completeness and none advancing.
5. Start where we bear the consequences
Our own finca is Site Zero. We try things on ourselves first, instrument our own systems, and live with our own mistakes before recommending anything.
Violation looks like: recommending a system we have never operated through a full season.
6. Cheap and behavioural before capital-intensive
The first questions on any site are “what is actually happening here?”, “what is being wasted?”, and “what changes if we simply operate it differently?” Equipment is one answer among several, and often not the first.
Violation looks like: a diagnosis that concludes in hardware every time.
7. Whole-system, whole-cost
Interventions are assessed on the full picture: capital cost, running cost, maintenance burden, lifespan, failure modes, and the effects they push onto other systems. A pool cover is a water intervention, an energy intervention, and a chemistry intervention at once.
Violation looks like: a payback calculation with no maintenance line.
8. Resilience and sustainability are two different goals
They frequently align, and sometimes conflict — a large backup generator buys resilience at a sustainability cost; a system tuned for minimum footprint may be fragile. We state which one an intervention serves, and name the trade-off rather than blurring it.
Violation looks like: the words used interchangeably in the same document.
9. Respect the place
Mallorca has its own climate, geology, water regime, building traditions, regulations and supply chain. Solutions imported unexamined from northern Europe or from generic sustainability literature are hypotheses, not answers, until validated locally.
Violation looks like: applying a heating strategy designed for a cold, dry, airtight climate to a damp stone building.
10. Write it down, date it, keep it alive
Knowledge that lives only in our heads cannot be shared, sold, audited or corrected. Every document carries a status, an owner and a last-reviewed date. Stale is acceptable; silently stale is not.
Violation looks like: this file, unreviewed, two years from now.