Biggest: largest, most ambitious, most lucrative, riskiest, highest impact? I have one that meets all of those criteria. Ditch legacy and redefine one language/OS to rule them all. Give app stores the middle finger. Throw the web browser away. Throw away antiquated build systems. Create a universal, minimal, extensible, and sensible layout and rendering engine. Create an imperative language as fast as c yet portable at the code level. Oh, and expect no one to invest in your madness. :) Be Steve Jobs and throw away the floppy disk and CD ROM before anyone else.
Look at the millions of complaints filed about HTML, JavaScript, CSS, etc. There is huge room for improvement in the way we write native and web applications - the fact that we still make a distinction signals its own problem. We can do better, and there will always be room for more innovation and improvement.
I agree. Similarly, keeping around tech with several decades of built-up cruft can also hurt your cause when you are trying to build an elegant, concise, and predictable system. :) You can keep putting a new coat of paint on 1950 Ferrari, even regularly changing out its parts, but it will never match something designed from the ground up with modern tech. The web browser, by nature of its spontaneous birth, ill-thought-out design, and insistence on backwards compatibility places many bent nails in the foundation. Adding on new layers to an old thing almost inevitably results in leaky abstraction.
I have projects that are 10 years old. The code isn't unusable, but I know I can do better. Sometimes adapting the old code just makes it ugly, hard to maintain, and error prone. So when it makes sense, I usually start fresh.
