In 2023, Bluebeam quietly stopped selling a native Mac version of Revu. The Mac product had never matched the Windows one feature for feature, and by the time Revu 20, the last perpetual licence most studios still cling to, reached end of support in 2026, the message was hard to misread. The centre of gravity in construction software is Windows, and the incumbent decided its Mac users were not worth the cost of a separate native build.

Plenty of architects read that as a betrayal. I run a small studio in Vilnius, and half of it works on Mac, so I felt the sting too. But I also build software, and from the builder's chair the decision looked less like an insult and more like a confession. Bluebeam told the truth about something most vendors hide: maintaining a truly native application on a second platform is expensive, it never stops being expensive, and at some point the spreadsheet wins the argument. That confession taught me more about building software than any conference talk ever has.

Native is a promise you keep every year, not once

When people say "native app" they usually picture a one-time achievement, a badge you earn at launch. It is not. Native is a standing debt. Every year the platform owner ships a new OS, a new chip, a new security model, a new signing requirement, and your app has to be dragged forward to meet all of it or it starts to rot in place. Apple's move from Intel to Apple Silicon is the clean example. It was a gift to users and a tax on every developer who had promised to run well on a Mac. You either rebuild for the new architecture or you ship something that limps.

Bluebeam had built its whole engine, its whole culture, its whole test infrastructure around Windows. Revu on Windows is a deep, serious tool. Studio Sessions, takeoffs, punch lists, RFI workflows, the machinery large firms actually run their projects on. To carry all of that to the Mac and keep it current, forever, against a user base that was a fraction of the Windows one, is a real cost with a thin return. So they stopped. As a business decision it is defensible. As a signal to the rest of us building tools, it is a warning: do not promise a platform you are not prepared to feed every single year.

The incumbent walked away from a segment, and that is information

When a market leader abandons a corner of its own market, that corner does not stop existing. The Mac architects did not vanish. Their heavy drawing sets did not get lighter. Their frustration got sharper, because the tool they had learned to tolerate now had a countdown on it. A vacancy opened, not because the incumbent failed, but because it chose focus over coverage.

I find that far more instructive than the usual startup story about a slow, lazy giant. Bluebeam is not lazy. It made a rational call to concentrate where its customers actually are. The lesson is not "the giant is weak, go attack it." The lesson is that even strong, well-run companies leave deliberate gaps, and a gap left on purpose by a smart competitor is often a better place to stand than a gap nobody has noticed yet. The purposeful gap comes with proof that people wanted the thing, and proof that serving them is hard. Hard is a moat if you are willing to pay for it.

Choosing where to build is the whole game

Here is the part that changed how I think. The most important engineering decision is not how you build. It is where you plant the thing you build.

About four in five of the people using Ncored are on Windows, which surprises nobody who works in architecture, engineering and construction. Windows is home. When we tune for the machine that actually sits under most of our users, we are tuning for an i7 or i9 with 32 GB of RAM, an RTX card and Windows 11 Pro, opening a 50-200 MB+ project set exported from AutoCAD, Revit and ArchiCAD. That is the reality we design against first, because that is the reality the work happens in. Pretending otherwise to look modern would be building for a brochure instead of a desk.

So the obvious, spreadsheet-approved move was to do exactly what Bluebeam did: go Windows, stay Windows, forget the Mac. We did not. We shipped an Apple Silicon build anyway, and I want to be precise about why, because "we love Mac users" is not a reason, it is a slogan.

Why we still shipped for the Mac

Three reasons, none of them sentimental.

  • Studios are mixed, and a tool that only serves part of a studio is a tool people quietly stop trusting. The technicians are on Windows towers, the principal is on a MacBook, and the drawing set has to look and behave the same on both sides of the table. A markup that only survives on one platform is worse than no shared tool at all.
  • The gap Bluebeam left was real, and I had lived inside it. I am not guessing about the Mac architect losing their perpetual licence to an end-of-support date. That is my own studio. Building for a pain you personally carry is a different act than building for a market segment on a slide.
  • Doing the hard version early is how you find out whether your engine is honest. If your speed only exists on one platform, it was never really your engine's speed, it was the platform doing the work. Making heavy CAD PDFs open fast and stay smooth on scroll, zoom and pan on a Windows workstation and on Apple Silicon forced us to earn the performance properly rather than borrow it. I will not get into how the engine does it, that part stays ours, but the discipline of two platforms kept us honest about what "fast" actually means.

None of that means we do everything, and I would rather say so plainly than let a nice essay imply otherwise. Ncored today is view, markup, text edit, page organisation, search, compression and layers. It does not do OCR, forms, e-sign or measurement yet. Those gaps are real, they carry a "yet", and pretending they are already closed would be exactly the kind of brochure-building I just complained about. If you want to see the performance claim tested rather than asserted, the numbers live on our CAD drawing benchmark page, not in this paragraph.

The real lesson: focus is a cost you pay somewhere

What Bluebeam leaving the Mac taught me is that focus is never free, it just moves the bill around. Focus on Windows only, and you save engineering effort and hand a segment to whoever wants it. Focus on both platforms, and you pay in complexity, test time and long nights making the same thing behave two ways. There is no version of this where you avoid the invoice. The only real choice is which bill you would rather pay, and whether the thing you are protecting is worth it.

Bluebeam decided its Windows depth was the thing worth protecting, and it protected it well. We decided that a mixed studio being able to open the same heavy drawing on any machine on the desk was the thing worth protecting, so we pay the two-platform tax on purpose. Neither of us is wrong. We are just answering a different question, which is the whole point. A big company retreating from a platform is not a scandal. It is a company telling you, out loud, what it has decided to be. The useful move is not to be offended by it. The useful move is to notice the gap it leaves, ask whether you are willing to pay the bill it walked away from, and if the answer is yes, go stand exactly there.