Anyone who liked this article may like to know that Alan Kay and his team at the Viewpoint Research Institute are currently making a silver bullet. A real one. Of the kind that really really hurts the Complexity Werewolf. They may not kill it, but the scars are already visible.
Take a look at their work, most notably the last STEPS progress report. They can use 50 times less code than well written, useful C code (like TCP). Compare with redundant, useless, badly written C++.
The secret is quite simple: you can write your own programming languages, and good code looks like a formal spec (to the point of being one, ideally). If you can't write good code, then write the language that will let you. How ? Start here : http://www.tinlizzie.org/ometa/
Other efforts similar in spirit but dramatically different in execution are Intentional Software and Jetbrains MPS. I really need to find a weekend to try to parse through VRI's work. Has anyone tried to make a digestable survey of the work they have done? I've glanced through their documentation but it seems quite sparse and their stuff is a moving target.
Kay and Simonyi are both PARC alumni, the place that has brought us a lot of the progress we've seen so far. This should be a slap in the face that current generations are making Angry Birds and Farmville, and these guys feel it is still upon them to do anything in the way of progress. It's just shameful.
I don't think they are similar in spirit. These other projects are missing the radical vision that is by far the most important and guiding thing behind Kay's group: to recreate all of personal computing in 20,000 lines of code. Operating system and applications. That is so profound that, if they succeed (a big if) it can't help but change the world. Simonyi and the Jetbrains people aren't in the same galaxy.
(I agree with you about it being kind of hard to follow what they're doing, though.)
I think you're missing the bigger picture here. It's not about a specific application, it's about the notion that the way we write software is fundamentally broken. All of these projects agree that each problem domain can potentially see orders of magnitude reduction in both LOC and bugs if custom languages are designed that fit the domain more appropriately than a one-size-fits-all approach that's been the status quo.
Take a look at their work, most notably the last STEPS progress report. They can use 50 times less code than well written, useful C code (like TCP). Compare with redundant, useless, badly written C++.
The secret is quite simple: you can write your own programming languages, and good code looks like a formal spec (to the point of being one, ideally). If you can't write good code, then write the language that will let you. How ? Start here : http://www.tinlizzie.org/ometa/