We're not using Google Spreadsheets as the model. Famo.us is open-source, which means an Opera user like yourself has the opportunity to make sure that your browser is not neglected.
We only work better on Webkit over other browsers because Webkit has the most mature implementation of the W3C APIs upon which we rely. We do not not rely on special vendor-specific features that are only implemented in Chrome or Safari.
If something doesn't work properly in some browsers, like Internet Explorer thus far, then that is because that browser has yet to implement the features upon which we depend or has yet to implement it correctly. We'll work to account for variability between implementations, but at the end of the day that's how things get done.
Native toolkits are battle-tested because they have had the luxury of time. There is nothing inherent about the web that excludes it from the opportunity of also being battle tested.
Our plan is not to have custom one-off UI components, but to actually mature what we're building into something that rivals what is done in native land. What better way long term is there for the open-source world to hedge risk against closed platforms than to move those native features to the web and perfect them.
"For things that are absolutely essential, like scrolling or
entering data, please use components I can trust."
On most sites I use, entering is only vaguely based on the native components with several layers of improvements achieved via HTML, CSS and JavaScript.
"I suppose it's only fitting that web devs would try, though to
be honest I hoped that impulse had ended with the death of
Flash."
Why would you hope that? Famo.us isn't the same as Flash in that it's open-source. The world now has a viable path to a solution that belongs to all of us on all platforms instead of being the domain of one company.
The only difference between the native components you keep on a pedestal and what we're building with the help of the community is time and polish.
Perform a mind experiment here: fast forward 1-2 years and just assume we are at parity with the native controls you prefer today. Given that what we're building is open and extensible to new unforeseen use cases, is that really a set-back or a leap forward in terms of what it enables us to explore as a community? The only thing standing between us and that future is time and effort.