The computer belongs to its user
A workstation should remain useful because its owner owns it. Local work should remain viable. Online services may extend the computer; they should not determine whether the computer works.
The rules Project Ceylon uses to decide what to build — and what it isn’t willing to build.
A workstation should remain useful because its owner owns it. Local work should remain viable. Online services may extend the computer; they should not determine whether the computer works.
Buying a legitimate physical product or software copy should mean the buyer owns that item or copy. Vendors may keep their copyright, trademarks, patents and other intellectual property in the underlying work — but continued use of an owned product shouldn't depend forever on a vendor keeping a licensing server alive.
A customer shouldn't need continuing vendor permission merely to use something they legitimately bought.
If a product needs a remote service in order to function, the vendor has a responsibility to provide a reasonable exit path if that service goes away.
This isn't being built just because it's technically interesting. The machine exists to help people get useful work done.
Multimedia should be first-class. So should ordinary business computing. Neither gets to be the afterthought.
Responsiveness is part of correctness. Fast hardware is not an excuse for inefficient software.
Internal sophistication doesn't require an incomprehensible user experience. Complexity should be organized and contained, not handed to the user.
Programming, graphics, music, video, automation, documents, presentations, games and design are central uses of the machine — not extras bolted onto it.
Owners and developers should have understandable APIs, scripting, command-line tools, documentation, examples and stable integration points.
Applications shouldn't have to live as isolated islands. The system should give software secure, documented ways to expose what it can do and work together.
Ceylon provides the platform. Developers keep ownership of their applications, engines, tools and intellectual property.
The order of preference is native capability first, then automation, then optional AI assistance. Applications have to remain useful when AI is unavailable or switched off.
Failures happen. A trustworthy system contains them, keeps useful evidence, explains what happened, and recovers where it practically can.
Don't declare success because something worked once. Test it. Measure it. Reproduce it. Document it. Review it. Test failure as well as success.
No exaggerated claims, no self-congratulation. Build the product, show what it does, publish the evidence, and let users, developers, reviewers and communities decide whether it deserves a recommendation.
None of this means we have all of those things finished.
We don't.
They're the rules we're using to decide what we're building — and, just as importantly, what we're not willing to build.
The journal is where you can watch us find out whether we can actually live up to them.