Constructionism seems to be one of Kay's sticking points -- this is one of the ways children can learn best (but what about adults?)
Obviously a modern Hypercard cannot simply be a contemporary reimplementation of Hypercard, with extended APIs for the online world etc. That, in fact, exists and it's called Livecode.
What was great about the later versions of Hypercard is how they fit holistically within OS8/9. You could interface with Applescript and access key APIs in the rest of the operating system in the plain way. What really should have happened next is that many of the basic applications should have themselves been implemented in "a hypercard" For example, imagine a Finder that let you peek under the hood to see all of its scripts and, after some kind of "unlocking" allowed you to change it. You get access to a lot of the power of the OS's API for free at that point, and users can easily figure out how to script something like a "every three days move files from here to there" or something.
The switch to Unix also meant a switch to and older, more complicated, less intuitive bundle of methods for system scripting. Useful UI metaphors go away at that level and are replaced with bad and outdated metaphors, like the teletype. Leaving aside things like Applescript which have been largely left to die on the vine, personal computing operating systems seem to have regressed from the contructionist perspective.
I guess what I'm suggesting is that you cannot have a "modern hypercard" without having a different kind of operating system, which itself might require a different hardware architecture.
An ideal system would be "layered." For example, at Layer 1 you have "a hypercard" and most of the GUI, including Finder-like things, windows, basic buttons, can be manipulated easily at this level. Users can also peek. But one they reach the limits of that high level hypercard system, then can "peel back" a layer and are introduced to lower level APIs and a new language. It has access to all the same objects and APIs as the hypercard layer, but in a different way, along with more APIs that were previously invisible. You could then have a third layer that is the systems language and APIs at the lowest user accessible level -- this is where "professionals" would live, but because it would be exporable and "peel back-able," determined users could get there by example.
Something like a lisp machine at layer 3 that runs a smalltalk at layer 2 that runs "a hypercard" at layer 1 would fit the bill. But take those as analogies.
I've come to a variant definition of constructionism, based not as a way to learn, but construction as a way to explore/understand. Especially when being entrepreneurial, analysis only takes you so far. In one of the YC video things someone said something to the effect that in Silicon Valley more people tend to talk about "what if you could?" rather than "why you can't", which sums it up for me.
The best example of a constructionist approach I can think of is complex adaptive systems (CAS)- the way to understand a CAS is to build one, taking one apart only gets you so far.
I see where you are going about the lisp and I like that! What a cool idea! lispercard? :-)
Obviously a modern Hypercard cannot simply be a contemporary reimplementation of Hypercard, with extended APIs for the online world etc. That, in fact, exists and it's called Livecode.
What was great about the later versions of Hypercard is how they fit holistically within OS8/9. You could interface with Applescript and access key APIs in the rest of the operating system in the plain way. What really should have happened next is that many of the basic applications should have themselves been implemented in "a hypercard" For example, imagine a Finder that let you peek under the hood to see all of its scripts and, after some kind of "unlocking" allowed you to change it. You get access to a lot of the power of the OS's API for free at that point, and users can easily figure out how to script something like a "every three days move files from here to there" or something.
The switch to Unix also meant a switch to and older, more complicated, less intuitive bundle of methods for system scripting. Useful UI metaphors go away at that level and are replaced with bad and outdated metaphors, like the teletype. Leaving aside things like Applescript which have been largely left to die on the vine, personal computing operating systems seem to have regressed from the contructionist perspective.
I guess what I'm suggesting is that you cannot have a "modern hypercard" without having a different kind of operating system, which itself might require a different hardware architecture.
An ideal system would be "layered." For example, at Layer 1 you have "a hypercard" and most of the GUI, including Finder-like things, windows, basic buttons, can be manipulated easily at this level. Users can also peek. But one they reach the limits of that high level hypercard system, then can "peel back" a layer and are introduced to lower level APIs and a new language. It has access to all the same objects and APIs as the hypercard layer, but in a different way, along with more APIs that were previously invisible. You could then have a third layer that is the systems language and APIs at the lowest user accessible level -- this is where "professionals" would live, but because it would be exporable and "peel back-able," determined users could get there by example.
Something like a lisp machine at layer 3 that runs a smalltalk at layer 2 that runs "a hypercard" at layer 1 would fit the bill. But take those as analogies.