consedge.
Blog

Edge Cases

Notes from the gap between a working demo and production software.

Written by the engineers doing the work.

Cadence
One technical article a month
Written by
The engineers on the work
Subjects
Revit · CAD/CAM · AI · QA
01What this is

Depth over frequency

Depth over frequency. One genuinely technical article a month beats four generic ones, and there is no penalty for publishing nothing in a slow month.

Every article here is written from work already done. If it cannot be, it does not go on this blog.

  • Revit & Autodesk
  • CAD/CAM
  • AI in AEC
  • QA & Releases
  • Practice notes
02Subjects

What the first articles cover

These are the subjects the first articles cover. If you need the answer to one of them before we publish it, ask — it is the kind of question a Feasibility Review exists for.

  1. Revit 2027 and .NET 10 — what plugin owners need to decide now

    One codebase with conditional compilation, or forks — and what each costs to maintain.

    Revit & Autodesk
  2. Why your Revit add-in breaks during a sync with central

    Transactions, sub-transactions and failure handling in a worksharing environment.

    Revit & Autodesk
  3. Testing a plugin across Revit 2022–2027 without testing everything

    How to prioritise a version matrix that has stopped being testable by hand.

    QA & Releases
  4. What an AI-generated Revit prototype actually costs to ship

    The distance between a demo that runs and software a team opens every morning.

    AI in AEC
  5. Deploying add-ins to 300 locked-down workstations

    Signing, antivirus, manifests, silent install, rollback.

    QA & Releases
  6. Deterministic rules or an LLM? Choosing for code compliance

    Why reproducibility decides the architecture.

    AI in AEC
  7. APS translation costs at production volume

    Caching and version-diffing, with the arithmetic.

    Practice notes
  8. “LOD 350” means four different things

    Writing LOD into a contract so handover does not become an argument.

    Practice notes

From prototype to production

AI made prototypes cheap. We make them ship.

A Feasibility Review is a 30-minute technical session with the engineers who would build it, followed by a written verdict: what the API actually allows, where it will break at scale, and what production costs. No obligation to proceed.

  • No charge, no obligation
  • Straight answers
  • Written verdict
  • Senior engineers, not account managers