Read-Only Digital Twins Before Greenhouse Automation
Why Streamline separates evidence reconstruction, prospective decision evaluation, and physical authority when developing greenhouse intelligence.
Editorial visualization · evidence, proposals, and future scenarios are labeled in the article.
Before new software receives authority over a greenhouse, it should be possible to inspect the evidence on which it would act.
That is the reason Streamline begins with recorded, read-only reconstruction. The development question is whether the system can preserve enough context to make an operating event understandable.
Read-only is a defined boundary: the workflow reconstructs evidence without issuing physical commands. It allows problems in identity, timestamps, calibration, and derived state to be investigated before they can become machine actions.
It is also useful on its own. A replay can show which records were available, where observations disagree, and what is missing. A team can use that information to improve capture practices or identify the next measurement needed.
The GH1 / Bed 4 evidence demonstrates one bounded instance of this work. The reconstruction preserved a mismatch between spatial observations and refused unsupported fusion. That result belongs to evidence engineering.
A recommendation introduces a separate question: does using this evidence lead to a better decision?
A proposed decision-support workflow should state the operating question, the information available at the time, the alternatives, the intended result, and the period over which that result will be assessed. Its evaluation should prevent later knowledge from leaking into an earlier decision.
Historical replay can help develop and challenge such a workflow. Prospective evaluation is needed to assess how it performs before the outcome is known. Operator usefulness, time spent reviewing, and the consequences of mistakes belong in that comparison.
Physical authority adds further requirements. Sensing must be adequate at the time of action. The task must be qualified for the equipment and conditions. Permission, interruption, observed completion, and recovery need to work in the actual environment.
Existing greenhouse controls already automate bounded processes. This sequence concerns the new intelligence and robotic capabilities Streamline proposes to add; it is not a claim that every established controller needs to be rebuilt around a digital twin.
Nor does passing reconstruction tests establish equipment safety. Software correctness, decision validity, physical safeguards, and operating economics answer different questions.
The operator framework addresses how responsibility could remain clear as authority grows. Human oversight has to be practical, with sufficient time and information to intervene where intervention is required.
A useful progression therefore has three distinct results: evidence someone can inspect, advice whose value has been measured, and physical work qualified within a specific scope. Progress at one stage can reveal that a proposed next step is unsuitable or uneconomic.
The current Streamline OS platform remains at recorded reconstruction. The broader biology-first architecture describes why that foundation could matter beyond the replay.
Read-only work earns its value by improving understanding today and making the requirements for additional authority explicit.
From the thesis to the operating system.
See what Streamline OS does today or bring us a greenhouse problem worth reconstructing.