top | item 12196513

Bret Victor, Inventing on Principle (2012) [video]

84 points| jacquesm | 9 years ago |youtube.com | reply

30 comments

order
[+] azeirah|9 years ago|reply
So, where are all the companies that are building next-generation developer tools and direct manipulation document and simulation tools? It has been a while since this video got released.
[+] pjmlp|9 years ago|reply
Apple has Swift Playgrounds as someone described.

Visual Studio has edit-and-continue, interactive REPL, visualizers, some sort of backtracking and now from Xamarin Interactive Workbooks.

Then you have the whole INotebook trend which started out with Python and nowadays supports multiple languages.

However all these tools are actually catching up with many of the features that Smalltalk-80, Interlip-D, Mesa/Cedar, Lisp Machines, Oberon already had.

This is what Bret Victor jokes about in another presentation of him, where he pretends we are in the 70's making predictions how the world of computers will look like in the 21st century.

The future of programming, https://www.youtube.com/watch?v=8pTEmbeENF4

[+] tolmasky|9 years ago|reply
We're building tonicdev.com (similar to iPython/Swift Playgrounds). A lot of these ideas are very difficult in the general case. For just one example, we spent a long time reasoning about time traveling (and came up with what we think is a great model for it - which takes into account the entire system including the filesystem, but it took considerable engineering effort and bleeding edge OS tools like CRIU: http://blog.tonicdev.com/2015/09/10/time-traveling-in-node.j... )
[+] amelius|9 years ago|reply
My guess: developer tools are being developed in open source now. It is not financially interesting for companies to build developer tools.
[+] doublerebel|9 years ago|reply
I built the PixiJS playground [1] to teach myself how (WebGL/Canvas interactive framework) Pixi works in a practical manner.

Many of the variables can be changed live in the REPL or with sliders as the code is running. Images can be dragged and dropped into the browser for sprites.

I built reactive (pre-React) Pixi components to make this possible. Victor's work is definitely an inspiration.

In the end I used the components and knowledge to build a touch-compatible shelf-configurator for a client selling shelving products.

[1]: http://pixijs-playground.nextorig.in

[+] fit2rule|9 years ago|reply
Take a look at the music-making industry, specifically all the synthesizers and DAW-style apps popping up on iOS. Those folks have been paying attention to Bret ..
[+] endergen|9 years ago|reply
Redux is the closest infrastructure that enables this stuff. The time travel aspects at least. Hot code reloading has definitely gone mainstream and if you look at a lot of the popular ones and trace their influences, many mention inspiration from Bret Victor.
[+] abcampbell|9 years ago|reply
Think you are missing (what I understand to be) his point.

The point is not that this is the UI of the future, the point is that there are other ways to conceptualize doing knowledge work with a machine, this is one.

Real innovation takes time...

Here's the man himself, Alan Kay, on the topic.

http://www.fastcodesign.com/3046437/5-steps-to-recreate-xero...

[+] hacker_9|9 years ago|reply
His demos only work with small amounts of code, scale up to the size of a real codebase and they all become infeasible, as well as not really that useful.
[+] pbadenski|9 years ago|reply
I’m developing Halik - time traveling debugger for JVM inspired by Learnable Programming. Emphasis on the ability to explore and visualize your program. See http://meet.halik.io.
[+] vincentriemer|9 years ago|reply
Using something like React-Hot-Loader & Redux-Devtools gets you pretty close for browser work.
[+] cma|9 years ago|reply
When Swift was first announced it showed off a lot of stuff that was inspired by it.
[+] nkrumm|9 years ago|reply
Are there any efforts do implement his "immediate feedback" ideas in the context of data science/scientific computing?

I could see some amazing iPython/jupyter demos, where the user can import some data, filter/rearrange data, and create a few plots-- but then adjust and see the plots change dynamically by adjusting values throughout the notebook.