Starting in 3D. It Took Me 24 Days Just to Launch the Software.

I have never once touched a piece of 3D software.

My whole life as a patternmaker has been flat. Paper and 2D CAD. I've checked volume in my head, in the toile, and in the finished garment. Standing a piece of clothing up inside a screen — I have zero experience with that.

A few years back I hit a fork in the road.
Bring in 3D CAD, or go dig into old clothes.

I didn't have the resources — money or time — to do both.
Learn the newest technology, or go exploring in the old and good.

I chose the latter.

The other day someone told me that what I'm doing with Fukuwohoru should sit well with 3D.
That advice planted a thought: maybe I should start poking at it, even a little at a time.

But putting time into something with nowhere to use it felt rough.
That's how I saw it — except I'd been mulling over the membership idea anyway, so I thought maybe I could work out the method inside that, and put out content that used it.

So this is that declaration piece, except honestly I haven't made anything yet.
The software just launched. And getting it to launch took 24 days.
Today I'll write about how that went, and why a flat-pattern guy past 40 is bothering to reach for 3D now.

Why 3D

The pattern data from the clothes I've taken apart for Fukuwohoru had, before I noticed, passed 50 patterns.

Taking a garment apart means breaking it to pull the structure out. You draw the thread, separate the parts, trace the lines. The garment dies, but the structure remains.
The DXF data sitting in the Library is that structure.

There's one thing that's nagged at me the whole time, though.

A flat pattern only stands up in the eyes of someone who can read it.

We patternmakers look at a pattern and go, "ah, so this is where it rises."
But that's the result of training, and most people simply can't see it. A 70-year-old structure I went to the trouble of tracing stays a cluster of lines, and the people it reaches stay few.

With a 3D simulator, I can take this pattern data and make the garment "stand up" again.

MEMO
3D simulator: software that reads pattern data part by part, and once you specify how the pieces are sewn together, it calculates the physical properties of the cloth and stands the garment up in three dimensions inside the screen. It's starting to be used in apparel planning; the one I've signed up for is a professional tool called VStitcher.

Deconstruction is breaking a garment to pull the structure out. 3D is that played in reverse. You stand the garment up from the structure.
The garment I broke returns, inside the screen, to a form you could wear again. As an exit for the work of digging, I thought this was worth doing.

What set it off was a single image I saw once. A look image that looked like nothing but a photograph turned out to use no photos at all — it was built from pattern data in 3D software. The person who made it said: "There's basically nobody with the nerve to learn this tedious work from scratch now."

If tedium is the barrier to entry, then this is easy. I'm the guy who's spent years piling up deconstruction — some of the most tedious work in the world.

And on top of that, 3D software can't have the machine understand which line gets sewn to which line.
A human has to specify each one, one at a time. Which means the most tedious part is exactly the patternmaker's domain. This is my ground, I thought.

If there's one problem, it's that 24 hours a day is given to everyone equally, and there's no way to make more of it.

And it won't launch

On August 16th, I signed up for a license at 75 dollars a month.

I installed it that same day, launched it, and got an error.

"VStitcher license isn't available for your account."

I mean, I just paid.

The admin screen said the contract was active. The seat showed as in use. And yet the software insisted it had no license.
I emailed support in English and got a form reply, then six days of silence. I chased it, a human finally appeared, and said: "This might be a bug where the license seat doesn't release. I've bumped you to two seats, so you should be able to get in."

I couldn't get in.

From here it was days of digging through logs with my AI partner. Same as taking a garment apart — you can only peel the cause away one layer at a time.

As I dug, the problem turned out to sit somewhere before the license. The login itself was never completing. On the login screen, even with the right password, the final bit of communication was getting turned away every time with a "400 Bad Request" error. Reinstalling the software completely made no difference. Support said, "maybe your security software," but when I checked, it was innocent.

In the end I cornered it with an experiment. I threw communication directly at the window that kept turning me away, changing the conditions as I went. And I found that if the package was small it got through, and if it was big it got rejected.

The culprit was cookies that had piled up in Chrome.

MEMO
Cookie: a small piece of data a browser stores per site. Used for things like remembering that you're logged in.

On my computer I've got a few tools I built myself running for other work, and the cookies they'd dropped in the browser had piled up to a bit over 4KB. The browser, dutifully, was sending every one of those cookies to VStitcher's login window too. And the window kept bouncing them back — "too much baggage" — along with the login notice that actually mattered.

I deleted the cookies and it was fixed in 10 seconds.

24 days, and 10 seconds.

It's like when a sewing machine is acting up and you suspect every part it has, only for the cause to be lint from the last job jammed in the hook.
The software wasn't broken. The license was fine. My desk was just a mess.

For the record, in fairness: when I sent support my analysis of the cause, they refunded the 24 days I couldn't use, right down to the day.

How I want to use it

This is counting chickens from a man who hasn't stood up a single garment yet, but let me put down what I'm thinking.

I want the clothes I've dug up to be something you can turn and look at. A garment I've taken apart can't be worn anymore. But the structure is all still there. I want to stand it up inside the screen so you can see it from the front, the back, the side. Things a flat, laid-out photo on a page can't convey — the way the seat curve escapes, that kind of thing — should become visible.

I want to line up the differences between eras as structure. This is a step beyond the numerical comparisons I do in the experiment room. Even the same 501 has different lines by era. The "difference in front rise" I put out as numbers — how does that come out as a silhouette? Fit them under the same conditions and compare, and it becomes a shape you can see with your eyes.

I want to test the "what ifs." A vintage original is one of one, so the experiment "what if this rise were 1cm deeper" is forever impossible. With data, it's possible. Build a hypothesis from real measurements, test it inside the screen, then go back to real measurements. I think it works as a tool for research.

There's one thing I've decided. I won't use 3D images in place of photos of the real thing.
In Fukuwohoru, the photograph is evidence. 3D is a research tool and nothing more; I have no intention of letting it pretend to be the real thing. I won't mix the two.

Last

I've picked the first garment. A 501 I took apart, known as the 501E. The work of writing out every seam line is, in fact, already done.
When I counted, there were 16 pairs. I'll sew these 16 pairs, one at a time, inside the screen.

It's my first time in 3D. The first garment will probably stand up awkwardly. A leg might bend off in a strange direction.

I'll show you that too, all of it, here.

24 days in, and I'm finally at the starting line.

I want to keep pushing forward a little at a time, in the gaps.

But right now I'm facing Fukuwohoru vol.6 with everything I've got, so it's a temporary ceasefire.