He Called My 900-Glyph Variable Typeface His Creative Vision — Then the IP Lawsuit Required My Foundry Signing Certificate
The drawing divider was a draftsman’s tool — two arms of steel, a brass hinge joint at the top, the tips sharp enough to mark paper. Mia Sundström used it to check optical spacing between letterforms on proof sheets, pressing the tips to two letters and measuring the visual gap.
The brass hinge joint was loose. It had been loose for four years, since the second year of her foundry, when she had dropped it on the studio floor and the hinge had shifted from its original tension. She had learned the specific angle that kept it set — slightly forward of vertical, maybe 5 degrees, a position she had internalized so completely that she no longer thought about finding it. Her hand went to the angle.
She was proofing the Voss variable font.
The commission had taken eighteen months. Lars Voss — Creative Director at the agency, managing the rebrand for the tech client — had brought the brief to her foundry in a PDF: a global technology brand, a new visual identity, a typeface that could work at display size on a billboard and at text size on a phone screen, with a variable weight axis that allowed the developer team to animate between light and heavy in real time.
She had designed the typeface.
The variable font was a 900-glyph system with five registered axes. The weight axis ran from 100 to 900 in continuous interpolation. The width axis ran from 75% to 125% of normal width. The contrast axis — which she had added after the second round of client review, when she realized the brand’s motion guidelines required a different optical behavior at small sizes than at display — had taken six weeks to calibrate alone.
The kerning table contained 14,000 individual pair rules.
She had written every rule by hand.
Pio was at the adjacent desk. He was 25, her design assistant, and he had drawn 200 of the 900 glyphs under her direction over the course of the commission. He had not drawn the kerning rules. No one had drawn the kerning rules except Mia, because kerning rules were not drawn — they were programmed, one by one, using the font design software’s script interface, and each rule required a decision about the optical relationship between two specific letterforms at a specific size in a specific weight range.
She was checking the A-V pair at 72pt in the Display weight — the heaviest weight, where the diagonal strokes of the capital A and the capital V were widest and the optical gap between them was most likely to look too tight.
She held the divider at the angle that kept the hinge set. She pressed the tips to the proof sheet.
Too tight. She adjusted the right sidebearing of the V by 4 units in the Display master.
She re-pulled the proof. She checked again.
Correct.
She made a note in the spacing file: V /Display / RSB +4u at 72pt and above.
The rule would apply automatically at all Display-weight settings in the final font file, calculated by the interpolation engine for the intermediate weights. This was the invisible layer of the work — decisions that no one would see unless they looked at the source file, and that the viewer would experience only as “the type looks right.”
She held the divider at the angle. She moved to the W-A pair.
The AIGA conference recording arrived in her inbox as a link, forwarded by a colleague with the message: “I think you should see this.”
She was at her foundry desk. She had been proofing a new glyph set — a diacritic extension for the Voss font, commissioned by the tech client’s European team six months after the rebrand launched. She set the divider down on the proof sheet, tips up, and opened the link.
The video was twelve minutes long. Lars Voss at the AIGA annual conference, keynote stage, blue blazer. The title on the screen behind him: “Designing with Variable Type.”
At minute three, he described the variable weight axis: “This is where I pushed the team hardest — the gradient from Light to Display had to feel effortless. I wanted the reader to feel the brand’s energy shift as the weight changed, and that meant engineering the interpolation to avoid the visual stutter you get when the curves don’t maintain their proportional relationship across the weight range.”
She had spent six weeks on the interpolation calibration. She had identified the visual stutter problem at week two, when the Light master’s curve nodes were misaligned with the Display master’s nodes at four specific junction points. She had fixed each one manually.
She watched the rest of the video.
At minute seven, he described “designing the variable axes to match the brand’s emotional range.”
She had designed the variable axes.
He had never opened a font file.
She watched the full twelve minutes. She closed the browser. She picked up the divider. The hinge settled to its angle in her hand. She went back to the diacritic glyph proof.
During the final client presentation, Lars had pointed to the variable weight axis on screen and said to the tech client’s CMO: “This is where I pushed the team hardest — the gradient from Light to Display had to feel effortless.”
Mia had been in the room. She had built that gradient across 14 interpolation masters over six weeks. She had said nothing.
After the meeting, Lars had said: “Great work. The client loved it.”
She had said: “The Display master needed an extra interpolation step.”
He had said: “Right. Smart.”
He had left.
She opened the font source file on the drive and made a note in the master spacing document about the Display interpolation step. The note would be there if anyone ever needed to update the typeface.
The brand press kit had been issued three days after the client presentation.
She had not been sent a copy. She had found it because Pio had a Google alert for type industry press releases and he had shown her the PDF on his phone in the studio. She had read the press kit.
It read: “The Voss Type System — custom typography developed by Lars Voss and the Creative Studio, designed to embody [Brand]’s commitment to human-centered technology.”
Her name was not in the press kit. The foundry name was not in the press kit. The phrase “Voss Type System” appeared fourteen times.
She had looked at the PDF on Pio’s phone for approximately five seconds. She had handed the phone back. She had picked up the divider. She had gone back to the proof she had been working on when he showed her.
Pio had said: “Should we —”
She had said: “No.”
She had meant: there was nothing to do about a press kit. A press kit was promotional material. It was not a legal document. It was not a certificate. The typeface files existed. The STF-2019-EU certificate was in all of them. The foundry registration was on file with the OpenType specification authority. The contract with Lars’s agency was in her project records, with her signature and the fee structure. None of that had changed because of the press kit.
She had kept working.
The 12-minute AIGA keynote video had come three months later. That was harder. Not because of the legal implications — which were the same as the press kit, none immediately — but because 400 type professionals in the audience had heard Lars describe designing a variable axis, and variable axes were not something you could describe from the outside. They were a technical architecture. You designed them or you didn’t. The audience had included people who knew the difference.
Some of them had looked at her in the hallway afterward.
She had answered their questions about the typeface directly, the same way she had answered the conference panel question: “I designed the typeface. He directed the project.”
That was accurate.
It was also insufficient. She knew it was insufficient. She had gone back to the foundry and worked on the diacritic extension until 11 PM.
At the type design conference, a colleague introduced her on the panel: “Mia Sundström — she designed the Voss rebrand typeface.”
Someone in the audience raised a hand: “I thought that was Lars Voss’s work? He spoke about it at AIGA.”
Mia said: “I designed the typeface. He directed the project.”
The moderator said: “Next question.”
She answered the next question. It was about variable font technology in editorial design, a topic she had been working in for ten years. She answered for four minutes. The audience took notes.
The email from Dana Cho arrived on a Thursday.
Subject: IP Dispute — [Tech Company] vs. [Rival] — Typeface Prior Art — Urgent.
Dana Cho was an IP lawyer at the firm retained by the tech company. The company was in a licensing dispute with a rival that claimed to have developed a similar variable font system independently and prior to the rebrand launch. The lawyers needed to establish prior art and prove originating ownership of the typeface — specifically, they needed the original font source files and kerning table data, with the type foundry’s digital signing certificate embedded in the file metadata.
She read: “digital signing certificate embedded in the file metadata.”
She opened the font source file on her terminal.
The file properties showed: Designer: Mia Sundström. Foundry: Sundström Type Foundry. Foundry Certificate: STF-2019-EU.
STF-2019-EU was the cryptographic signing certificate she had obtained for her foundry in 2019 from the OpenType specification authority — a unique certificate tied to her foundry identity, embedded at export into every .TTF and .OTF file her foundry produced. It was cryptographic proof of origin. It could not be transferred. It could not be fabricated. Every font file she had exported for the Voss commission — and there were 47 of them, across 5 axes and multiple test builds — carried STF-2019-EU in the metadata.
Lars’s agency had no type foundry signing certificate. They had never exported a font file under their own identity. They had no STF number. They had the font files the tech client was using, but the metadata in those files said Sundström Type Foundry.
She looked at the divider on the proof sheet. The tips were up, the way she had left them when she set the tool down to open the email. The hinge was settled at its angle on the paper.
She closed the font file. She did not call Lars. She checked Pio’s schedule — he had three glyphs left to finish on the diacritic extension before she could do the final kerning pass. She wrote him a note about the order she wanted them in. She went back to the proof.
Lars read the IP dispute brief in a meeting with his agency’s legal team. He was confident. The client had the font files. The font files were the typeface. The typeface was the creative work of the agency. He told the legal team to pull the file inventory and prepare the IP filing.
The legal team pulled the files.
They read the metadata.
STF-2019-EU. Sundström Type Foundry. Designer: Mia Sundström.
They called him back.
She had been working on the diacritic extension for six weeks.
The tech client’s European team had contacted her directly — not through Lars’s agency — with a request for a diacritic extension to cover Western and Eastern European character sets. She had estimated the work, sent a quote to the client, and begun the design after the contract was signed.
Lars’s agency was not part of the diacritic extension contract. She had not told Lars about it. The tech client had come to her because the original typeface files were hers — the metadata said Sundström Type Foundry, and any type professional who looked at the files would see where to go for extension work.
This was how the type industry worked. The foundry that produced the original certification was the foundry you contacted for derivative work.
She had added 47 new glyphs to the original 900. Each new glyph required spacing decisions — how much space on the left side, how much on the right, how the glyph related optically to its neighbors in the character set. The diacritics required additional kerning rules: the ä-v pair, the ë-n pair, the ö-a pair, each with different optical properties from their base letter pairs because the diacritic mark shifted the visual center of the letter.
She had added approximately 400 kerning pair rules for the diacritic set.
She had not told Lars about those either.
The divider was on the proof sheet when the IP email arrived. She picked it up afterward and continued.
The legal team explained the problem in five minutes.
A type foundry signing certificate — specifically, an STF-series OpenType certification — was cryptographic proof of the originating type foundry. The certificate was unique to the foundry, generated from the foundry’s registered identity and private key, and embedded into every font file the foundry exported. It was not a visible label. It was not a metadata field that could be changed by a client or an agency. It was in the binary structure of the font file.
The rival’s lawyers would challenge the IP filing. When they examined the font files, they would see STF-2019-EU. They would ask who held that certificate. Lars could not produce it. His agency had never had a type foundry signing certificate.
More specifically: in an IP dispute over a typeface, the originating designer was the foundry that produced the certified files. Under IP law and OpenType licensing standards, the foundry certificate established the originating designer in the legal record.
The certificate holder was Sundström Type Foundry.
Lars looked up Sundström Type Foundry’s registration: Mia Sundström, sole proprietor, registered 2019.
He thought about “designing the variable axes to match the brand’s emotional range.” He had said it at the AIGA conference, on the keynote stage, in front of 400 type designers and creative professionals. He did not know what a variable axis was, technically. He had heard Mia use the phrase “variable axis” in a client feedback session and he had used it in the keynote because it sounded precise and correct.
He was now in a position where a rival’s IP lawyers would ask him, under oath, what a variable axis was. He would not be able to answer. The STF-2019-EU certificate would answer instead.
He called Mia.
She was at the proof table when the call came. She had the diacritic extension laid out — 47 new glyphs, the cedilla variants, the ring-above characters, the combined diacritics that required new spacing rules. She had the divider in her hand, checking the ë spacing against the base e.
He said: “The lawyers need the foundry certificate and your appearance as the originating designer in the IP filing.”
She said: “I know.”
He said: “STF-2019-EU is your certificate.”
She said: “Yes.”
He paused. He said: “I should have understood the attribution implications from the start.”
She said: “Yes.”
He waited. She did not add more.
He said: “The filing needs to be submitted by Friday.”
She said: “I’ll have the certificate documentation to Dana Cho by Thursday.”
She put the phone down. She picked up the divider. She held the hinge at the angle that kept it set. She checked the ë spacing again. The distance was correct. She marked it on the proof and moved to the next pair.
She had known from the beginning that the certificate was in the files.
Not as a strategic decision. As a technical fact. Every font file produced by a registered type foundry carried the foundry’s signing certificate in its binary metadata. This was the OpenType standard — a format designed to allow font files to be distributed across the internet while maintaining an authenticated chain of origin. She had obtained the STF-2019-EU certificate when she registered the foundry in 2019, before the Voss commission, before she had any major client. She had obtained it because it was part of operating a legitimate type foundry, the same way a contractor obtained a professional license.
She had not put the certificate in the Voss font files as proof against future misattribution. She had put it in the Voss font files because she put it in every font file she produced.
The certificate had been there for 18 months, embedded in 47 files, while Lars had given conference talks about designing the variable axes.
She thought about this for approximately three seconds on the afternoon the IP email arrived, while she was looking at the divider on the proof sheet. Then she wrote Pio a note about the glyph order and went back to work.
She did not feel vindicated. The certificate was not vindication. It was a record. It had always been a record.
She sent the STF-2019-EU certificate documentation to Dana Cho on Thursday morning.
The package included: the Sundström Type Foundry registration (2019, EU OpenType authority, sole proprietor Mia Sundström), the certificate chain documentation showing STF-2019-EU as the issuing authority for all 47 font files exported for the Voss commission, and a signed declaration confirming she was the designer and that the typeface — including the variable axis architecture and the 14,000-pair kerning table — was her original work.
The IP filing was submitted Friday. The record read: Type system designed by Mia Sundström, Sundström Type Foundry, STF-2019-EU. Creative direction by Lars Voss, Creative Studio.
The rival’s claim collapsed when their lawyers examined the STF-2019-EU certificate chain. Every file in the typeface system had been produced by Sundström Type Foundry. The rival had no equivalent documentation. Their claim of independent development could not be sustained against a cryptographic origin record.
The settlement was reached in six weeks.
Dana Cho’s email arrived Monday of the seventh week.
“Ms. Sundström — your foundry certificate was the clearest proof of origin I have encountered in a type dispute. The rival’s claim has been dismissed. You will be cited as the designer of record in the settlement documentation. The tech company has also agreed to update all brand documentation to reflect the correct originating designer attribution.”
Pio heard about it from Mia that afternoon, in the studio.
He said: “Your certificate. Your 14,000 pairs.”
She said: “Yes.”
He said: “Worth counting.”
She said: “I counted them.”
She had counted them. Every one. Because you could not write 14,000 kerning pair rules without counting them. The rules were in a spreadsheet with 14,000 rows, one per pair, each row containing the pair name, the pair spacing value, and the weight range it applied to. She had maintained the spreadsheet throughout the commission. It was referenced in the STF certificate package as the technical basis for the kerning documentation.
Lars called after the settlement closed.
“Good outcome. Your foundry certificate was decisive.”
She said: “The signing key is embedded at export.”
He said: “Right. I should have understood the attribution implications from the start. I want to make sure the press correction is handled properly — you’ll be listed as the originating designer in all brand communications going forward.”
She said: “Yes.”
He said: “Great work, Mia.”
She said: “Thank you.”
She hung up. She picked up the drawing divider. She was back at the proof table, the diacritic extension laid out in front of her. She had twelve pairs left to check.
The AIGA video still had 28,000 views. The description had been updated: “Type designed by Mia Sundström, Sundström Type Foundry.” Most people did not read video descriptions.
She knew this.
She held the hinge at the angle. She moved to the next pair. The spacing was right. She marked it on the proof.
The IP settlement documentation was filed in the court record as a permanent legal document.
Under IP law, the settlement established Mia Sundström, Sundström Type Foundry, as the originating designer of the typeface system. The filing was public record. The tech company’s IP counsel had cc’d her on every filing document, which she had filed in the Voss project archive folder — seven documents total, from the initial IP filing through the settlement agreement.
The rival company’s lawyers had examined all 47 font files. They had found STF-2019-EU in every one. Their review had taken four days. After four days, they had advised their client that the independent development claim could not be sustained against the certificate chain.
The rival had conceded.
Dana Cho had sent her a note outside the formal correspondence: “For what it’s worth — I have been doing IP work in the type industry for twelve years and I have never seen a cleaner origin record. The certificate chain, the kerning documentation, the axis registration — every piece of evidence points the same direction. There was never a question.”
Mia had read “there was never a question.”
She had filed the note in the same archive folder as the settlement documents.
She opened the diacritic extension proof. She had twelve pairs left. She picked up the divider. She held the hinge at the angle. She went back to work.
The YouTube video was there. 28,000 views. The description had been updated. Most people did not read video descriptions. She knew this. She had known it since the day she first saw the video, and she had gone back to the diacritic proof then too.
She picked up the drawing divider and held the brass hinge joint at the angle that kept it set — slightly forward of vertical, the angle she had found by accident in year two and never forgotten. The steel points rested on the proof sheet at exactly the distance between the ‘A’ and ‘V’ at 72pt in the new typeface. The optical spacing looked tight. She adjusted the hinge one millimeter. She tested again. The spacing was correct.
The IP settlement document was in the project archive — “Mia Sundström, Sundström Type Foundry, STF-2019-EU, designer of record.” Dana Cho had cc’d her on the final filing. Pio was at the adjacent desk, drawing the lowercase numerals for the new commission. The YouTube video still had 28,000 views. She had stopped checking. She held the hinge at the right angle. The spacing was correct. She marked it on the proof and moved to the next pair.
The new commission was a magazine typeface — a serif, which she had not designed in three years, for a European editorial group. The brief was specific: optimized for long-form reading at 10pt on coated stock, with a companion mono variant for captions. She had two axis proposals in draft and a glyph count estimate of 720 for the base roman.
The client had asked about variable technology. She had explained variable fonts in practical terms: what they could do for the editorial workflow, what they could not, where the design decisions became technical decisions that affected the reader’s experience invisibly. The client had listened carefully. They had asked good questions.
The commission had been won on the strength of the tech rebrand work, which the editorial director had mentioned by name. He had said: “We want the quality of the Sundström typeface.” Not the Voss Type System. The Sundström typeface. The press correction had reached the right people.
She had started the serif skeleton sketches in January. It was now March. She had 47 sketches, the beginning of a spacing system, and a plan for the optical size axis that she would present to the client in two weeks.
She was proofing the ‘a’ and ‘e’ lowercase in the current stage of the design. The ‘a’ had a double-story construction — her default for text faces — and the aperture needed to be slightly wider than she had initially drawn it, to maintain readability at 10pt.
She pressed the divider tips to the proof sheet.
The aperture was correct now.
She marked it. She moved to the ‘g’.
The Voss commission brand press correction had arrived two weeks after the settlement. She had read it. She had filed it in the Voss project folder. The press correction read: “Designed by Mia Sundström, Sundström Type Foundry, commissioned by Creative Studio.” Her name was listed first. She had filed it.
The 28,000 YouTube views were still there. She had stopped counting them the day the settlement was filed. The type design community had seen the conference panel, where she had answered the question directly. That was enough.
She held the divider at the right angle. The lowercase ‘g’ aperture was correct on the first check.
She moved to the next letter.
