It started as procrastination. I had a set of comments due back to a structural engineer, the drawing was sitting on my desktop, and instead of marking it up I found myself wondering something I had been wondering for years: is it just me, or does every PDF app fall apart on a real construction sheet?

So I ran a small experiment. One drawing. One machine. Nine apps. I would open the same file in each, one after another, and pay attention not to a stopwatch but to the thing you actually feel: the moment between double-click and being able to work.

The setup

The file was a coordination sheet exported from our BIM model, the kind of thing that lands in my inbox every week. In our studio the drawings we push around are AutoCAD, Revit and ArchiCAD exports, and a full 50-200 MB+ project set is normal, not the exception. This was one dense sheet from that world: tens of thousands of separate line and text objects, hatches, embedded fonts, the usual invisible clutter that CAD exports carry around like packaging.

The machine was deliberately ordinary for an architecture office. A desktop with an i7, 32 GB of RAM, Windows 11 Pro, a mid-range RTX card. Nothing exotic, nothing starved. If a PDF app struggled here, it was not going to be the hardware's fault. I lined up the nine apps most of us actually reach for on Windows: the browsers already on the machine, the free minimalist reader everyone recommends, a couple of the big general editors, a budget editor with a loyal following, and a few others. I will not turn this into a scoreboard with numbers next to names, partly because a single afternoon on one machine is a story, not a lab result, and partly because what mattered was never the number.

Watching them open

The first thing I noticed is that almost everything opens the file. That sounds like a low bar, and it is, but it is also the source of a lot of confusion. When someone tells you their PDF app is fine, they usually mean it opened. Opening is the easy part.

The browsers were the surprise of the day. Double-click, and the sheet appeared quickly, faster than I expected. For about four seconds I thought the whole experiment was going to be anticlimactic. Then I tried to zoom into a stair detail, and the drawing turned to soup. Pan a little, wait for the tiles to redraw. Zoom again, wait. It was like reading a newspaper through a keyhole that moved half a second after my hand did. Fine for a glance. Useless for a review.

The free minimalist reader was the opposite personality. It threw the sheet on screen almost instantly and stayed quick when I panned, and for a moment I understood why people swear by it. Then I went to leave a comment, which is the entire reason I had opened the file, and remembered why it lives on the "reading" side of my brain and not the "working" side. It was built to show you a document, not to let you argue with one.

The big general editors were the most interesting to watch, because you could see them thinking. The window would appear, the toolbars would populate, and then there was this pause, a held breath, while the app digested the page. Ten seconds. Fifteen. Longer than I wanted to stand there. And once it finally settled, every zoom carried a faint drag, as if the app was doing the whole sheet's worth of work again each time I moved. These are powerful, serious tools. On a heavy drawing they felt like driving a moving truck through a car park: enormous capability, wrong shape for the job.

The budget editor did better than its price suggests and worse than its fans promise. Quick enough to open, competent to mark up, and then it hit the same wall the others did once I pushed into the densest corner of the sheet, because that corner is where a general-purpose renderer meets a file that was never meant to be read by a human in one piece.

The thing the exercise actually revealed

By the sixth or seventh app I stopped caring about opening speed, and that is the real finding. The interesting axis was not "how many seconds to open." It was "can I do my job while it is open."

There is a gap, and once you see it you cannot unsee it, between an app that opens the file and an app that is usable on the file. Opening is a one-time cost you pay at the start. Usable is a running cost you pay on every single interaction for the next hour: every zoom into a connection detail, every pan across a plan, every jump between sheets, every markup you place and then move because you put it in the wrong spot. An app can win the opening race and then quietly tax you for the rest of the afternoon. Several of the nine did exactly that.

This is why "it opens fine for me" is such a slippery claim. The person saying it opened the file, nodded, and closed it. They never spent forty minutes zooming into a stair core while a colleague read numbers off the screen. The friction does not live in the opening. It lives in the work.

Why I stopped waiting for one of them to be the answer

I should say the obvious thing plainly: I did not run this experiment as a neutral observer. I built one of the nine. Ncored exists because I got tired of exactly this afternoon repeating itself in our studio, and I wanted a tool that treated a heavy drawing as the normal case rather than an edge case it apologises for. So when I tell you it opened fast and stayed smooth to scroll, zoom and pan while I marked the sheet up, read that with the appropriate pinch of salt and go re-run the test yourself. I put the reproducible version, same idea, your files, your machine, on a benchmark page you can work through, because the point of the exercise is that you should not take my word for the one app I am attached to.

I will also be square about what it does not do. Today it is view, markup, text edits, page handling, search, compression and layers. It does not do OCR, forms, e-sign or on-drawing measurement yet, and that "yet" is doing real work, not covering for a permanent gap. If your review depends on takeoffs and measured quantities right now, some of the heavier general editors still hold ground I do not, and I would rather tell you that than pretend otherwise.

What I would tell a colleague

If you take one thing from my wasted-then-not-wasted afternoon, make it this: test the tool on your worst real drawing, not a friendly sample, and do not test the opening. Open it, then work. Zoom into the messiest detail. Pan across the whole sheet. Place a markup, move it, delete it. Jump between pages. The apps that survive that are the ones worth your money, on Windows or on Mac. The number on the stopwatch was never the story. The half-second of drag on every single move, multiplied by an afternoon, is the story. Nine apps taught me that opening a drawing and being able to use it are two completely different promises, and most tools only make the first one.