Galatea

My user experience needed an overhaul. I had just finished a class on computer vision, and was feeling confident enough to create an iOS app that scans your face, superimposes mask designs on you on real time, and then lets buy the mask right in the app. I've finished writing the simulation part of the app -- now we just need to get details right.

Wrapped my 2D editor in a React web app within a month and removed most features except the most basic, robust ones. Integrated Stripe and linked it to my landing page. Wrote a few posts about the page in neighborhood groups and posted a few pictures on Instagram. Because I knew no seamstresses, I decided not to assemble the masks myself -- I sold them as easy sew kits. Spoiler: it did not work out too well. I sold four mask kits, and the person had a great deal of trouble putting them together. People did not seem to use the best features at all -- finding a ruler and measuring their faces was a task most people were far too lazy to do. Furthermore, I simply did not reach enough people that were willing the sew the masks on their own.

Half an hour of tweaking allowed me to repurpose my custom garment 2D editor into a custom mask 2D editor. Masks are highly in demand, easier to rapidly prototype and iterate on, and are in their infancy in terms of styles and trends. I decided to focus on making higher end, stylish masks with mix-and-match ties.

After finishing the 2D editor, I started showing it to people. It soon became obvious that nobody knows what clothing looks like before its sewn (except for sewing enthusiasts and fashion designers). Without a 3D preview of the garment, a 2D flat pattern editor utterly confuses people and hinders their creativity. I decided to learn 3D graphics programming, and create features that allows users to sculpt a garment on a 3D model just by dragging and dropping their mouse. The value of this editor is that it doesn't take any special skill to make sure the outfit looks nice -- there are constraints on the sculpting of the garment to ensure continuity, smoothness, and symmetry (optionally).

I compiled all of the research, feedback, and ideas from the past few years and pulled it together into a long term vision for the project. I took a good look at emerging technologies, market trends, and societal changes that could potentially impact the development of a bespoke tailoring services. Using all of this information, I created a detailed explanation of product requirements.

For example, after studying the social media landscape, I decided that Galatea should be a maker platform and community, rather than just an editor that you use by yourself. I thought about the "remix culture" exhibited in Reddit and Twitter memes, TikTok clips, and TinkerCad's 3D design platform. Part of what makes TikTok and TinkerCAD so accessible and rich is the ability to create and edit in the same environment that you consume the content. I envision Galatea as being similarly community and content based, where you can scroll through other people's virtual designs, make "remixes" or "spin offs" of other artists' designs, republish your version, and have the ability to get likes, comments, and feedback. You can order any publicly shared design in your size, kind of like how you can download and 3D print any community designs on TinkerCad.

Of course, so much has changed since the original writing of the roadmap, namely 3D scanning and COVID-19. I'll discuss this in a bit.

Over a year and a half, I spent a few hours a week after school to make a 2D web editor for garments. With every slice, cut, and distortion of the fabric, seam relationships and markings are automatically updated. It also made the process of procedurally generating designs inside of garments easy and intuitive.

When using Adobe Illustrator for fashion design became tedious, I decided to learn graphics programming and write software that helped me design clothing. As I needed more features, I just programmed them for myself. I drew some inspiration from the formulas I learned at the NY fashion Academy, but I also tried to optimize operations for a virtual design environment (instead of a physical environment where design is done with tracing paper, muslin, pencils, pins, and dressforms). Eventually, I was able to design clothing that fit me perfectly quickly, precisely, and efficiently! After some time, I realized that in my software, I could swap out my measurements for someone else's. Now, anyone can design their own clothing! That is, if I manage to make the user interface comprehensible to anyone besides myself. This turned out to be a very challenging task, but came with the opportunity of turning my software into a business.

I began using machines at the makerspaces at the University of Washington to make elaborate, precisely constructed garments more efficiently. Through using the laser cutter, computerized embroidery machine, and printers, I was able to automate a lot of the construction and also discover techniques that are not taught at the New York Fashion Academy. I also started to heavily integrate Adobe Illustrator into the workflow, by using it to create so called engineered prints, and measure seam lengths. While Illustrator worked great for the first dozen prototype garments, I eventually decided that Illustrator was not powerful enough to automate tasks like matching seam length and parameterizing the measurements.

I wanted to learn about garment sizing and construction. While I continued experimenting on my own, I also wanted to learn industry norms and have the opportunity to engage with people in the fashion community. I started taking classes at the New York Fashion Academy (in Ballard, near Seattle, WA). I took classes like Sewing I, II, and III, as well as Patternmaking I, II, III, and Sloper. I still keep in touch with my instructor, Candace, for mentorship and advice -- she now runs the Seattle Fashion Academy in Lake City.

Question: Can I create apparel that is better suited to by tastes?

Exploration and experimentation: I spent a month becoming familiar with fabric and stitching. I cut, ripped, pulled, ironed, washed, and stitched many different types of fabric to see how different material behaved under different conditions. I also learned how to use a Brother sewing machine and studied how different settings for stitch styles, stitch lengths, stitch width, and thread tension would affect the appearance of seams.

I also discovered Spoonflower, a service that prints image files onto fabric and delivers the fabric to your home. Seeing this as a tool to help me customize apparel to my tastes, I began free-handing outlines of garments in Illustrator and filling them with graphics. I wanted the designs to flow continuously across the garment, rather than being cut off at seams like I often saw with garments. I printed garments and sewed them intuitively, not taking too much time to learn proper stitching and finishing techniques.

I read lots of sewing forums and tutorials to see how garments are constructed. The instructions were complex and were dense with jargon that I was unfamiliar with. However, it was helpful to see the garment creation process from sewing hobbyists.

Takeaways: From online resources, I affirmed that it was indeed possible to make garments without a factory. To my surprise, I learned that some of the most elegant and professional looking garments could only be made by hand. However, inexperienced home sewists rarely produced work that looked professional. From this, I learned that it usually took years or even decades to learn to avoid amateur sewing mistakes.

The most important mistake revealed by my initial trials was the sizing. I estimated that my free-hand sketched garments would reach my mid-thigh. I learned that clothing looks much wider when unwrapped and laid out in a 2D plane because it must wrap around the sides of your body. The garment was ill-fitting in many other ways -- the armholes were too large and angled outwards, and the garment was even more shapeless than I had intended. My next step was to find the simplest way to create a garment that fit me.