Creating BUSY

In 2019, I began my journey with generative algorithms and pen plotters. It started when I discovered fascinating examples of it on Twitter's #PlotterTwitter hashtag on a Hacker News post. Artists worldwide were experimenting with drawing the output of code with machines. There was a lot to absorb. Some built DIY plotters, others restored vintage machines from the 1980s, and others purchased Axidraws. After a few days of pondering whether to buy one, I decided I would, but only if I could make the machine pay for itself by selling drawings online. An Axidraw A3 was soon on its way.

I downloaded Processing and got to work. After a few months, I switched to P5.JS and the p5.js.svg plugin to generate my vector graphics. By early 2020, I worked hard on an algorithm called "A Series of Tubes." A few things were at play: multiple ink colors, fun details, and infinite results via randomness. I eventually sold a limited number of these drawings on Etsy at $135 a piece, and my Axidraw did indeed pay for itself. I posted on #plottertwitter, too, which got me on the radar of Casey Reas, who approached me to do a group exhibition focused on plotter art with Feral File with five other contemporary artists working with plotted artwork.

Fast-forward to January 2024

After a string of successful exhibitions, my passion for generative artwork led me to quit my corporate IT job and focus on making artwork full-time. Evil Mad Scientist, the makers of the Axidraw, released a large-format A1 edition that doubled the working area of the model I had. It was a no-brainer that I needed one, so I placed another order.

I wanted to challenge myself to utilize the machine to its fullest capacity. How much detail and complexity could I cram into a single piece of paper? Code makes this relatively easy. Making it interesting was my artistic challenge.

I felt my prior work with A Series of Tubes was never fully realized, and I frankly didn't have a strong desire to work within its primitive code base. Instead, I decided to make a spiritual successor that did many of the same tricks but also made more. That's when I dreamed up the concept of BUSY.

Within generative art, there can be a challenge of creating relatable things. Often, you will see artists expressing abstract or mathematical concepts to significant effect. I decided that a human element was essential and sought out a way to incorporate hand-drawn elements into the generative systems. I created a pipeline that accepted SVGs and converted them into manipulable geometry, thus allowing me to draw and insert buildings, cars, and other elements. This flow was a tremendous leverage point; the more illustrations I created, the richer the art became, and it was not dependent on complex functionality that I had to write continuously. The code would integrate the new data, dispersing it across the system and manipulating its color, size, and orientation.

Wimmelbilders had inspired me in the past. They're highly complex drawings, often with hidden meanings, that force the viewer to sit with the art for extended periods to appreciate all of the detail. Artists such as Richard Scarry, and nostalgic book series such as "Where's Waldo?" are fine examples of this style. It occurred to me that it may be possible to replicate this effect using computation and pen plotters.

To achieve a highly dynamic and multi-layered system, I had to crack the nut of geometric occlusion. At some point in every plotter artist's career, they realize that overlapping shapes is quite tricky. Typically, the pen plotter will happily draw the occluded geometry since SVGs do not actually clip stacked elements. This challenge makes it hard to create artwork with depth. Zancan gave me an overview of his solution that he used to great effect on "Garden, Monoliths" (2021). I built my own version for "ORI-Plotted" (2023) but found my implementation's performance lacking since it used image buffers to check canvas occupancy.

I rolled my own solution using haphazard logic of my own intuition. It almost always worked, except when it didn't. This time, it was a math-based solution instead of a buffer-based one. I discovered that performing tens of thousands of occlusion operations in one composition would create all sorts of edge cases that had to be dealt with. In time, I solved nearly every issue. Some solutions were quite hacky, like moving every polygon's vertex by a minuscule, random amount. There's no better feeling than the moment you finally solve a tricky logic problem after spending countless hours mired in it.

My thoughts on physical media were evolving at this time, too. I adopted fountain pens as my primary tool for pen plotting. A couple of things spurred this. I realized that archival-quality ink was a baseline requirement for all artworks from now on, as my goal was to make art that would stand the test of time. If I'm being honest, too, I had become enamored by the community surrounding fountain pens on Reddit. Reviewing the posts about new inks and pens had become very enjoyable. Fountain pens came with other advantages compared to my previous choice of fineliners. Since the nib was metal and the ink was easily refillable, I never needed to replace pens.

I developed a workflow to expand BUSY's variety of tubes and objects. I created a programmatic switch that allowed me to focus on a single type of tube segment and iteratively alter its appearance. As I had done with the illustrations, I kept dreaming up new ways to render the tubes and add them to the algorithm. Some were inspired by sorting and shuffling algorithms, a callback to some of the earliest plotted drawings that caught my attention and featured visualizations of sorting algorithms.

To maintain a simple working system, I established a rule that each pipe would be comprised of segments, and each segment would accept an array of ink colors and return a new array. This process allowed me to chain the segments together and have continuous lines throughout the pipes. I also developed the segments in one orientation. I used transformations to rotate and flip them - a significant improvement from my prior work, where I calculated everything in four separate functions for each possible direction.

Scaling up production

From the beginning, my goal was to make each BUSY drawing stand alone as a premiere piece of art and do a large production run. In the past, my editions of plotter artworks have been 30 in total. I wanted to exceed this, partially because it sounded challenging and kind of insane. I was ready and excited to commit to 500-700 hours of drawing time.

Making 1-of-1 plotter artworks with some trial and error is not particularly hard. Any issues, whether related to registration or ink flow, can be written off as the "personality" of the art. Seeing such aberrations is a significant reason I enjoy plotting artwork; it becomes imperfect in the most extraordinary way.

This cavalier attitude was not my intention for BUSY, though. It would not be productive to recreate drawings due to production mishaps. Instead, I needed to define a process that led to almost all artworks being completed in one attempt with high fidelity.

My first challenge was registering each of the 5-12 separate ink layers embedded in each drawing. I could mostly accomplish this feat using a common clamp point on each pen. Disaster would occur if, for any reason, I needed to pause, refill, and resume a layer. It was impossible to keep everything in a perfectly static location.

In conversations with other artists, I learned that some had begun using 3D Printing to augment their practices. This was the next opportunity for me as well. I bought a printer and taught myself the basics of Autodesk Fusion. With a lot of trial and error, I designed a custom pen holder and retention system for my fountain pens that solved all of my issues. I could clamp and unclamp any number of pen bodies and have assurance that my lines would be tightly aligned when their nibs hit the paper.

Practically speaking, drawing an iteration of the BUSY algorithm is a time-consuming and hands-on process. The pen plotter must be reloaded with a new ink color every hour or two. If it sits idle for too long, the pen will dry up and refuse to draw until thoroughly cleaned. These requirements have incentivized me to be present during each drawing and to have an accurate idea of each layer's duration.

Out of the box, Axidraw pen plotters require a hardline connection to a computer to operate. They will provide some time estimates during plotting. If I needed to go out to a meeting, I couldn't bring my laptop, or the drawing would stop until I returned.

My focus turned to optimizing the processes of controlling the plotter and planning its starts and stops. First, I purchased a NUC, a small form factor computer that runs Linux. I developed custom software for this computer that managed my pen plotter queue. It could simulate each color layer of a drawing and give me an estimate of drawing time. Using a tool called Vpype, it also optimized the pen plotter's path, minimizing the pen-up movements and connecting lines that begin and end at the same exact location. When I started a new pass, the software posted information to custom clocks and progress bars that are physically located in my art studio and gave me simple visual queues of its progress. This way, I was ready to plan a day around my pen swaps and drawing schedule.

Hyperlocal Generative Art

I began to consider BUSY more of a framework than a single artwork. Perhaps it could be refactored into a system that mimicked a specific location. Since I had developed a pipeline to ingest illustrations, what if I illustrated the landmarks of real-world places instead of dreaming them up?

There was a fine line to draw - I never saw myself adding things like the Eiffel Tower to the main BUSY algorithm because it felt inauthentic and less interesting than my creations. However, I was intrigued by modeling artwork after a specific location and live-plotting it on location. Maybe I thought this was a great way to do a world tour as an artist; I'm not sure.

An opportunity came to trial this idea in Marfa, Texas, for the yearly Artblocks Weekend, attracting artists, collectors, and curators worldwide. For around 1-2 months before the event in November 2024, I drove around the desolate high-desert town in western Texas on Google Streetview. I had visited Marfa once before in 2022 and had many ideas of places to draw, and with the help of locals and Google, I soon built a collection of around 50 landmarks from around town. I tried to get all the big things, like the iconic water tower and town hall, along with some abandoned buildings and small shops. Marfa is not a busy place, though, so I wrote some code to generate large pastures and desert expanses. BUSY In Marfa was born, and you can see more about it here.

I borrowed a Nextdraw pen plotter from a fellow artist who lived in Texas and set up a mock art studio on location. Every day, I methodically plotted each layer of my drawings. Sometimes, I was in the location at 9-10 PM at night, doing a final pass to be ready for the next day. Seeing a plotter work in real life is mesmerizing, and throughout the extended weekend, attendees came by and could see the process in person.

A magical moment occurred when someone asked me if they could use my artwork to inspire a tattoo they were getting on the spot. Of course, I agreed, and the illustration of the Marfa water tower that I did ended up on someone's forearm.

Really big drawings, please

In the last few years, I've become friends with the creators of Axidraw and Nextdraw. Originally a company named Evil Mad Scientist, they were acquired by Bantam Tools in 2024 and relocated from the West Coast to Peekskill, NY, to produce machines. I've been continuously impressed by their business's focus on artist advocacy and happily attended the first Plotterfest three-day event in the summer of 2024.

I proposed the idea of doing really big plots to Bantam Tools and was excited to hear that there was some possibility of a prototype machine being built in early 2025. Amazing things would happen if we could somehow coordinate the launch of BUSY with a limited production window on a new large-format machine.

Thankfully, I had built my algorithm in a way that allowed me to work off of a fixed density level per square inch. To create an artwork of a different size, I simply had to change its set size, and the same relative scale would exist in relation to the existing 22"x30" paper that I had chosen for the run of 50 BUSY drawings.

Just days before the public announcement of BUSY, I was in Peekskill, assessing whether it would be feasible to create five large-format drawings. It was questionable whether we could keep the same level of detail with the dramatically larger plotter size, and I was told that increasing the size of every component in engineering rarely just works. It was an experiment to find the limit of what was possible, and as it turned out, we could draw up to 72" artworks. Another big consideration was time and distance; just because it was possible didn't mean it was practical. With some calculations and tests, I landed on using 60"x24" sheets of paper and using 4 ink colors. All in all, these larger drawings would need around 16 hours each for production.

Time was running out, and I was 300 miles from home. Bre Pettis, owner of Bantam Tools, kindly loaned me his minivan so I could borrow the prototype and transport it to my studio in Vermont. Obviously, building and lending me an extremely expensive one-of-a-kind prototype machine is incredibly generous, while also letting me use his family minivan for a week was almost too much.

My house can barely fit one Axidraw A1 pen plotter. It certainly was not ready for a machine with a footprint of nearly 8'x2.5'. My wife agreed to let me take over our dining room so that I could set the prototype up. I even used the minivan to go buy some 4'x8' MDF from Home Depot, which served as a table. Anyone who has had to transport sheet products from a hardware store without a right-sized vehicle can relate to the struggle and perhaps the relief of gaining access to a minivan.

For a week, I ran the plotter all day and into the night and produced five 5'x2' drawings. These would be entitled "BUSIEST" and exemplify the concept of the artwork in grand form.

The prototype's next stop was the Bitforms Gallery in New York City for a BUSY opening party. I wanted to showcase the machine in action by drawing a 5'x2' artwork on location and to maximize the impact, and I wanted it to finish at the end of the party. There were several logistical hurdles to overcome because the plot would need to run for 12 hours before the final step, which would occur during the event.

We hatched a plan to start the drawing in Vermont, complete it 75% of the way, and then transport it in place down to Manhattan. If I could pull this off, it would be 🪄 a magic trick, given how little room for error there is with the alignment of the four colors and the 300-mile drive in the back of a minivan.

We arrived at Bitforms at 2 PM and began staging the art and display for the prototype pen plotter. Using very rough guesswork, I determined a start time for drawing around 4 PM. I homed the machine, inserted the Pilot Kakuno fountain pen into my 3D Printed holder, and pressed start. The machine picked up where it left off with no issues. It finished the artwork right as the event concluded that night.

What's next

From the outset, I made a series of decisions with one goal in mind: Ensuring that this artwork will be viewable in digital and physical format for many decades to come. That's why I insisted on it being committed to the Ethereum blockchain in its entirety and why I chose archival inks and paper. Just the act of transcribing the results of a generative artwork to physical media detaches it from the risks of digital archiving. I have a high degree of confidence that BUSY stands the test of time.

BUSY is currently being exhibited on Artblocks.io. The release is occurring in three phases:

• Auction for 5 BUSIEST drawings (update: sold out in 8 minutes)
• Auction for 50 BUSY drawings (taking place from 4/8, 12PM EST to 4/9, 12PM EST with up to one hour of extensions)
• Prints of two selected outputs from BUSY, going on sale on 4/14

To see more about the project:
View the Exhibiton here
Watch the wonderful launch video here