There's a term I use, demoware, which I use to specifically refer to things that are easy to demo, but hard to convert into real systems.
Another observation I need to set this point up is that there are three numbers in computer programming: 0, 1, and "many". In particular, once you've done the work to generalize from 1 to 2, you might as well just generalize from 1 to n. But this is often a huge leap.
So, combining the two points: It's demoware to create a single program that can accept multiple input sources and do something sensible with them. It's a STONKING ENORMOUS change to go back over the entire past 45 years of computing history and ask absolutely everything, from the kernel drivers up to the end-user programs, to generalize from a 1-source-of-input model to an n-sources-of-input model. Even beyond the mere driver programming and the mere task of getting the OS layer and beyond to handle this, you're doing things like asking every program to include UI-level locking and change conflict resolution policies, which is an incredible degree of complication to add to things. (Text editors have some moderately simple answer to this, that doesn't particularly generalize.) It isn't even as simple as pushing the capabilities down to the OS level, you need massive modifications to end-user programs as well.
That's why this was ignored for 45 years. If someone just handed you a system that did this on a silver platter, you might find it fantastic to use, but the multiplicative cost of trying to make this work on the systems of the past would be prohibitive; what was easy to write in a demo is freakishly expensive when multiplied out over all the relevant code bases.
(Note I am aware that some progress has been made on this front, with multiple USB keyboards and such, that Linux distro that runs 4 XWindows servers off of one system, etc. Obviously this is still not exactly pervasive functionality, though.)
I think that is exactly the issue. John Underkoffler, who invented the interface for Minority Report and thought it was realistic, has been making collaborative digital platforms since at least 1999 http://web.mit.edu/invent/iow/underkoffler.html and took it commercial in 2006 http://oblong.com/
Comments
There's a term I use, demoware, which I use to specifically refer to things that are easy to demo, but hard to convert into real systems.
Another observation I need to set this point up is that there are three numbers in computer programming: 0, 1, and "many". In particular, once you've done the work to generalize from 1 to 2, you might as well just generalize from 1 to n. But this is often a huge leap.
So, combining the two points: It's demoware to create a single program that can accept multiple input sources and do something sensible with them. It's a STONKING ENORMOUS change to go back over the entire past 45 years of computing history and ask absolutely everything, from the kernel drivers up to the end-user programs, to generalize from a 1-source-of-input model to an n-sources-of-input model. Even beyond the mere driver programming and the mere task of getting the OS layer and beyond to handle this, you're doing things like asking every program to include UI-level locking and change conflict resolution policies, which is an incredible degree of complication to add to things. (Text editors have some moderately simple answer to this, that doesn't particularly generalize.) It isn't even as simple as pushing the capabilities down to the OS level, you need massive modifications to end-user programs as well.
That's why this was ignored for 45 years. If someone just handed you a system that did this on a silver platter, you might find it fantastic to use, but the multiplicative cost of trying to make this work on the systems of the past would be prohibitive; what was easy to write in a demo is freakishly expensive when multiplied out over all the relevant code bases.
(Note I am aware that some progress has been made on this front, with multiple USB keyboards and such, that Linux distro that runs 4 XWindows servers off of one system, etc. Obviously this is still not exactly pervasive functionality, though.)
I think that is exactly the issue. John Underkoffler, who invented the interface for Minority Report and thought it was realistic, has been making collaborative digital platforms since at least 1999 http://web.mit.edu/invent/iow/underkoffler.html and took it commercial in 2006 http://oblong.com/