There's plenty of new ideas around, you don't have to generate them yourself. If you ask in forums or search the web you can get a big list. The problem is culling the list for practical and successful ideas. Somebody has to invest the time and/or money do actually do it, and accept the high risk of failure typical of startups and new open-source projects. That's probably the bottleneck.
Ideas are cheap, execution is not.
Here's a free sample tip: Dynamic Relational. It's the new "plastic" (a movie reference). The NoSql movement has shown a market/desire for dynamic databases. But with Dynamic Relational you get dynamism AND sql; you don't have to choose one or the other. And you can set constraints/rules to gradually make it more "static" like traditional RDBMS.
I feel, in my opinion, getting an idea and generating something comparatively easier part compared to actually educating the market / users about what the idea is and what is the solution we have generated, we all see / are a part of so many good products which just had a steep learning curve and hence failed / didnt get accepted.
I'm curious of specific examples. Functional programming is possibly an example. It takes a good long time for many to become productive with it, and with some it will never click. It's hard to actually know if its a personal preference or something inherently better. After all, code is more for humans to read than it is for computers.
Ideas are cheap, execution is not.
Here's a free sample tip: Dynamic Relational. It's the new "plastic" (a movie reference). The NoSql movement has shown a market/desire for dynamic databases. But with Dynamic Relational you get dynamism AND sql; you don't have to choose one or the other. And you can set constraints/rules to gradually make it more "static" like traditional RDBMS.