Hot or not: subtext programming language

As a real programmer I cringe at the Polymorphism example, only because I see the point of objects is to encapsulate data with the associated methods and this breaks all of that. Indeed you try to avoid inheritance and use composition instead because of the very problem exposed. These examples struck me as a deeply procedural low level way to represent these concepts, although that common simplicity does suggest that really all objects are is a way to do if statements better. That in itself is a idea that everyone knew but few people ever said.

I find the GUI editor with a mouse painful and this GUI isn't helping! With a heavily user optimised interface with at the very least some of the smarts of Excel and a lot of hot keys (get rid of the dialogs, stop making me right click to splice etc) then it has some potential.

What I think we need is this approach translated into a real language with real libraries that have side effects and all the lovely mess that is a real programming environment. Right now its hard to see just how bad it is for the other 99% of programming that is grabbing from a DB and throwing data on a screen (those poor corporate drones!).

How about the author hosts his website with the language! Building enough infrastructure to do that is bound to help advance the language to the point where its more than just a toy.

An interesting simple way to consider logic will be interesting to watch this evolve.

It would be hot if anything new was happening with it, saying that the video gives away enough of the concept that there is nothing stopping someone from implementing it themselves. I personally think the dude should go open source with it but I also understand the pain of releasing your child into the wild before you think it's ready!

If someone can implement a useful programming language with non-textual structure and do it right, that would be pretty hot. (There are a few out there, but they really need better editor support...)

I have had something similar in mind for a while, but when I saw Subtext, I found that what I had been thinking about program representation (as inspired by git) was very similar to the stuff in Subtext's papers, except that Subtext's was much more developed in the vein of Smalltalk. On the other hand, I found the decision tree stuff to be something of a curiosity; I'm not sure if it's necessary for the Smalltalky functionality.

Subtext looks interesting.

I think this type of visualization would be fantastic as an eclipse plugin for Java. It would be at a minimum a great teaching tool to new programmers.

No "real programmer" would ever admit to using it, but I'd love to get an implementation that could talk to an existing set of libraries. Looks like an innovative way to get some advantages of strong typing and such while inverting the learning curve.