2ndOpinion.FYIEN
genius.wiki

#552 2012 · Figma (Dylan Field, Evan Wallace) · Software / design tools

Figma didn't add real-time collaboration to a design tool — it rebuilt the rendering engine so collaboration stopped being a feature and became unavoidable

问题

A feature that competitors treat as a UI add-on can be structurally impossible to bolt onto an existing product if the product's underlying architecture was never built to support it, no matter how much the team wants to ship it

背景

By the early 2010s, professional design software — Photoshop, Sketch, and their predecessors — had always run as native desktop applications working on local files, because that was the only architecture capable of rendering complex vector and raster graphics fast enough to feel responsive. That architecture made real-time multi-user editing, the way Google Docs had already made word processing collaborative, structurally unavailable: a desktop app editing a local file has no natural mechanism for two people to see each other's cursor and edits appear live, the way a shared cloud document does.

Dylan Field and Evan Wallace, who met studying computer science at Brown University, saw a different possibility in April 2011 when Wallace built a real-time fluid simulation, called Water, that ran entirely inside a browser tab using WebGL — a graphics API built for 3D applications, with no plugin or download required. Rather than start by designing collaboration features, they spent years building a browser-based rendering engine using WebGL and a C++ engine compiled to WebAssembly capable of hardware-accelerated, professional-grade 2D vector graphics — something no one had proven possible for a serious design tool at the time.

换别人会怎么做

Add real-time collaboration as a new feature to an existing desktop design tool -- build a sync layer, a live-cursor system, a conflict-resolution mechanism on top of the existing local-file architecture -- the standard product-development approach of layering a wanted feature onto whatever foundation already exists.

他们看到了什么

Field and Wallace saw that collaboration wasn't missing from Photoshop and Sketch because nobody had built the feature, it was missing because a desktop app editing a local file has no natural mechanism for two people to see the same live state at all -- the architecture itself made the feature structurally impossible, not just unbuilt. The fix wasn't a smarter feature, it was a different foundation where shared live state was the default rather than something to be added.

那一手

Because Figma's design surface lived inside a browser tab connected to a server by default, rather than a file sitting on one person's hard drive, live multi-user editing followed almost automatically once the rendering engine existed: any browser tab open to the same document was already talking to the same shared state, so showing a second person's cursor and edits in real time was an extension of the architecture rather than a separate system bolted onto it. Figma's public preview launched in December 2015, and its full real-time collaborative editor shipped in September 2016 — collaboration wasn't the headline engineering problem the company solved, it was the byproduct of having already solved a harder one.

为什么管用

By building the rendering engine itself to run inside a browser tab talking to a server, every open tab on a document was already connected to the same shared state by construction, so showing another person's cursor and edits live was a natural extension of how the system already worked, not a separate system that had to be designed, synced and debugged on top of an architecture built for single-user local files. Because the hard, multi-year engineering problem (GPU-accelerated professional graphics rendering in a browser) was solved first, the feature everyone actually wanted arrived nearly for free once the foundation existed, and competitors retrofitting collaboration onto desktop-file architectures could only ever produce a more limited version of the same capability.

值了多少

Figma grew from a two-person hackathon proof-of-concept into design software used by roughly 13 million users by 2025, the year the company went public on the NYSE. Shares opened at $85 on July 31, 2025 and closed the first trading day at more than triple the IPO price, giving the company a market value of $56.3 billion, one of the largest software IPO debuts of the decade — competitors including Adobe, whose 2022 attempt to acquire Figma outright for $20 billion was blocked by regulators, never matched Figma's browser-native collaborative architecture with a comparable retrofit of their own desktop tools.

什么时候会失灵

This approach requires being willing to spend years building foundational infrastructure with no visible feature to show for it, a bet only viable with patient capital and a team confident the architectural rebuild will eventually pay off, since most organizations and investors expect faster, visible progress than a from-scratch rendering engine can deliver. It also depends on the desired capability genuinely requiring a different foundation rather than just better engineering at the existing layer -- rebuilding an entire architecture is enormously expensive, and misdiagnosing a solvable feature-layer problem as an architecture-layer one wastes years pursuing a rebuild the original stack could have supported all along.

后来呢

Figma's browser-first, GPU-accelerated architecture became the reference example in software and product-design writing for how real-time collaboration in creative tools has to be earned at the rendering-engine layer rather than added as a feature, and rival design and productivity tools that tried to add live multiplayer editing onto existing desktop-file architectures generally shipped slower, more limited versions of the same capability years later.

资料来源

  1. [1]FigmaWikipedia, 2026en.wikipedia.org
  2. [2]How Figma's Dylan Field Won Over the HatersSequoia Capital, 2021sequoiacap.com
  3. [3]FigmaEvan Wallace (Figma co-founder and former CTO), 2021madebyevan.com

收下它

同一路聪明