How We Work
This page is the difference between us and “take the job and start coding”: delivery is split into stages with exit criteria, and if the evidence is incomplete at any stage, we do not move to the next one.
Four steps and their exit criteria
Requirements
→ Process description, hardware model and order number, I/O conditions, safety level, acceptance criteria
Acceptance criteria must be verifiable assertions — not phrases like “logic must be correct”
Solution frame
→ POU structure planning, variable naming conventions, state machine design, deliverables list
Block call depth limited to 3; safety logic separated from standard logic
Clarification
→ Open items logged, each with both branches (if yes / if no)
Only written evidence closes an item (report / nameplate / datasheet) — verbal confirmation does not count
Implementation & delivery
→ Code and self-check report → real compilation → static analysis → dynamic verification → on-site confirmation
If the previous stage fails, the next stage is not even evaluated; safety functions must be verified independently and signed
Three hard rules
What we need from you
- Requirements: process needs, I/O conditions, safety level, acceptance criteria
- Solution frame: POU planning, naming conventions, deliverables list
- Clarification: open items logged with both branches of action
- Implementation: stage-gated — evidence complete before moving on