BeOS
In 1990, Jean-Louis Gassée and Steve Sakoman left Apple to found Be Inc. Their project was to create an entirely new operating system capable of taking advantage of modern machines without dragging along the burden of backward compatibility. Apple was going through a difficult period at the time, with its aging system showing its limitations in the face of hardware advances.
BeOS was born from this desire to start from scratch. Be’s engineers built a system designed primarily for multimedia, without worrying about the usual compromises. Multithreading runs through all layers of the system, with each task divided into multiple execution threads to make the best use of resources. The 64-bit BFS file system incorporates database capabilities and journaling to protect data. Symmetric multiprocessor management proved refined, which meant that BeOS ran faster on a Pentium 233 with 64 MB of memory than Mac OS X on machines eight times more powerful.
The 1995-1996 demonstrations were impressive. BeOS simultaneously ran multiple real-time 3D rendering windows, played video, and browsed the web without faltering, where competing systems struggled. Apple, desperately seeking a modern replacement for its aging system, entered into discussions to acquire Be Inc. in 1996. The negotiations failed over price. Apple turned to NeXT, bringing Steve Jobs back to the company he had created.
Be Inc. continued on its path and targeted the Intel PC market starting in 1998, while keeping its PowerPC version. The system won people over with its elegant interface, responsiveness, and technical innovations like dynamic queries in the file system. Users appreciated the stability, execution speed, and fine-grained file attribute management.
But BeOS failed to break through. Applications were sorely lacking. Developers hesitated to invest time in a platform with few users, and users waited for applications before adopting the system. A vicious circle. Windows dominated the PC market, Mac loyalists stayed on their platform, and space was tight for a newcomer.
Be Inc. attempted a pivot in 1999 toward emerging Internet appliances. Too late. Investors grew weary of funding a project that couldn’t find its audience. The company closed in 2001 and sold its patents to Palm.
The story doesn’t end there completely. Haiku, an open source project, took up the torch and reconstructed BeOS in spirit. Some of the system’s ideas—advanced metadata management or multimedia-oriented architecture—spread elsewhere.
BeOS illustrates the difficulties of a market where technical excellence is not enough. Timing plays its role: BeOS probably arrived too late to disrupt established systems. What remains is the mark of a modern architecture, an attention to performance that continues to inspire. The system shows that by freeing oneself from historical constraints, remarkable results can be achieved.
The lesson concerns the software ecosystem. Without popular applications, a system, however brilliant, withers away. This reality still weighs on current strategies, where building a developer community matters as much as technical prowess.
BeOS’s commercial failure does not erase its technical contribution or the questions it raised about market dynamics.