Working across
the whole problem.
I build complete systems out of ambiguous, half-understood problems. Technology is the tool. The work is finding what actually needs solving, then proving the answer works.
See the work ↘01 / Selected projects
Each started with a concrete problem and changed as I learnt more.
A / 01
Live / in production
A social product built and tested through real events
Hulofuse started with a difficult question: can software do more than introduce people, and actually help them form friendships? We built it to find out. It launched across the Bay Area, grew to more than 800 users, and brought people together at real events.
See how the production system works ↗The decisionThe second meeting mattered more than the first.
A / 02
Complete / live
An animation community rebuilt from the server up
Artists animate an 11-second audio clip, vote, and receive professional critique. The job was to move the neglected old site onto modern infrastructure. The server turned out to be the easy part. The operating system, application, forums, media pipeline, spam controls, and deployment all needed work, so the migration became a complete rebuild. A local Qwen classifier and vision checks now catch different forms of spam. The web platform is live, and its mobile companion lets members stream 16 years of animation entries, vote, and use the forums.
Read why the migration became a rebuild ↗The decisionRebuilding became simpler than preserving each failing layer.
A / 03
In use / still developing
A file-sync replacement that grew around the work
Fogjaw began by replacing a file-sync client that used a small website running on the same computer to manage local files. Solving that exposed everything that happened around the file: review, organisation, delivery, and time. A few friends now use it in their studio work while it continues to develop.
Read how the software followed the work ↗The patternSolving the first problem revealed the system around it.
A / 04
In active use / evolving
Reconstructing systems nobody can explain
Lazarus takes an inherited system nobody can fully explain and builds a verified source of truth from what survived and what is actually running. It checks documentation against repositories, build systems, hosts, processes, jobs, configuration, logs, and dependencies, then uses that evidence to plan and build the replacement. It is used regularly and continues to evolve when a real system exposes a new case.
Read how it replaces guesses with evidence ↗The decisionThe running system was the only reliable record.
02 / Currently building
NextTang is an independent, open-source port of the ZX Spectrum Next FPGA design to Sipeed Tang Console boards, starting with the 138K model.
The work is at the honest early stage: prove the board first, then adapt the core. Clocks, video, memory, storage, audio, and input each need reproducible bring-up before a compatibility claim means anything. There is no working bitstream yet.
Each stage needs evidence on the actual board before the next one counts.
03 / Workbench
Games, tools, and working systems that started with curiosity, an irritation, or something I wanted to exist.
01
Rebuilding a childhood game for the ZX Spectrum Next exposed gaps in the available knowledge and tooling. I added a developer debugger to the JNext emulator, built source-level tools, and began extracting the repeated game systems into a reusable engine.
Ongoing / working tools
02
A collaborative SSH application built so several people can work in the same remote terminal without passing control back and forth. It is complete and already used by a lot of people. The source is private for now.
Complete / in use
03
A complete mobile portal for Animation Mentor. Students can take their classes, access course material, and use the school’s social features from the app. It is live and in use.
Complete
04
My reimagining of what my favourite ZX Spectrum game would look like in 3D if it came out today. It is a complete game with online multiplayer. Making it meant working across game design, modelling, rigging, sound, networking, and visual polish.
Complete / online multiplayer
DRAWER / 01
05
Recorded family conversations become cited transcripts, a memory graph, and the next question worth asking. The imported-audio loop works end to end.
Working end to end
06
A phone-based automotive diagnostic tool for reading faults, streaming live vehicle data, inspecting freeze frames, checking emissions readiness, and identifying the vehicle through an OBD adapter.
Working / live vehicle data
07
A complete strategy trainer with three drill systems. It ranks all 15 discard choices by expected value, explains pegging decisions, and shows when board position changes the best play.
Complete
08
An iPhone and Apple Watch breathing app that keeps animation, audio, haptics, and session timing in sync. The current build installs and runs on a device.
Working / iPhone + Watch
09
A Linux port for running Wallpaper Engine assets across Wayland and X11, with multi-screen placement, property controls, and frame-rate limits.
Working / Wayland + X11
10
Typing recipes into the Breville Control Freak one by one was tedious, so I reverse-engineered the encryption used by its recipe files and built a macOS and Windows manager to read and write them. Then I modified a 16GB SanDisk Connect Wireless Stick and rebuilt the Wi-Fi app running on it. The manager could connect while the stick stayed plugged into the cooker, so I could update recipes from my phone without carrying the stick back and forth.
Complete
04 / About
I’m Jon Attree. I grew up working class in England. As a child, I was always taking things apart to find out how they worked, though they rarely still worked once I’d put them back together.
Some work starts with an inherited system nobody wants to touch. Some starts with an irritating tool, a product that misses the human problem, or a game I wanted to make as a child. I work across product, infrastructure, data, operations, and the parts that do not fit neatly inside one discipline.
01Start with the human problem.
02Test claims that sound too easy.
03Build the missing tool.
04Let evidence change the plan.
05Leave a working system behind.
05 / Contact
If you’re sitting on a strange idea, a system nobody wants to touch, or a problem that doesn’t fit neatly inside one job title, I’d like to hear about it.
Start a conversation