Cruncher: An Implementation of Bret Victor's Scrubbing Calculator | Hacker News Reader

Cruncher: An Implementation of Bret Victor's Scrubbing Calculator

cruncher.io

112 pointsluu32 comments

In case you're unfamiliar with the term "scrubbing calculator" (like me), read this: http://worrydream.com/ScrubbingCalculator/

It really is a good essay, but the videos in it have seem to be broken for awhile.

They work in Firefox

I'd blame Chrome :P. It worked not so long ago. Also, it works if you right-click and then select "Open video in new tab".

In chrome, I had to click on the words "click to play" to get the video running.

It didn't work for me (using chrome latest).

Author's other projects are pretty interesting too: https://rsnous.com/

Very cool! Thanks for sharing!

When I click on the "assorted examples" link, it pops up a dialog to ask me if I want to leave the page.

After I tell it that yes, I really meant to click on the link that it had offered me, it displays the examples... and promptly replaces them with an empty page after a few seconds.

That annoying dialog also comes out when doing other things, like using the Back button.

Interesting!

Quick feedback: some I believe issues in the current version, in the "assorted examples":

- "4 apples + 5 oranges * 18 friends" ― wasn't that meant to be "(4 apples + 5 oranges) * 18 friends"?

- "Hold ALT while scrubbing to plot" ― doesn't seem to work for me on Fx nor Chrome

- after clicking on "assorted examples", they like to not show immediately, requiring me to go back till the URL has "?/examples/cdg" and then refresh.

- also, Ctrl-Z doesn't seem to do anything, I'd really appreciate if Undo was available after I click something by accident.

- 5,000 in some locales means 5 not 5000 (',' is used as decimal separator instead of '.') ― see e.g.: https://en.wikipedia.org/wiki/Decimal_separator#Countries_us...

> "4 apples + 5 oranges * 18 friends" ― wasn't that meant to be "(4 apples + 5 oranges) * 18 friends"?

In the 'real world', that would clearly be the correct interpretation, even if it flies in the face of the 'standard' rules of precedence. If the calculator really works that way, it only makes it more impressive in my book.

Why is it that way in the real world? The example is still abstract, and without clear interpretation without context.

The text is supposed to be, in the first place, all the relevant context in an abstract form, i.e. problem minus all irrelevancies.

> Why is it that way in the real world?

It's an interesting discussion how the nature of concrete items involved in calculations might affect the precedence. I think that, since "apples" and "oranges" are clearly more closely related than "friends", it makes sense that they bind together more closely in this case. Also, the order in which the items are presented has some bearing on the interpretation.

Sounds like DWIM, NWIS...

On the one hand I love that people are thinking about how to make algebra more approachable and more usable in everyday life, but on the other hand I fear that it might mess with the learning curve. Once you want to do something that's beyond the capabilities of the scrubbing calculator, you start from zero because nobody's ever told you about algebraic equations and unknowns and you only know how to describe your problem in terms of "connections" and "locked numbers."

Compare this to an application like Calca (http://calca.io/). Calca is more difficult to start with for a mathophobe because, while it will solve your calculations and equations for you, you do have to explicitly define all variables and formulate an equation. But the result is that the learning curve when you graduate from the application is much smoother: you can easily port more complex problems to a programming language or pen-and-paper math. Your existing knowledge doesn't suddenly become void.

The only problem there a bridging this to systems that can solve more complex cases. Might be difficult, I admit; Maybe introduce these concepts/UIs (connections etc) into math apps?

One of my co-workers introduced me to Dragonbox the other day. It purports to teach algebra through a game. (There are two versions; one for 5 year olds and one for 13 year olds.)

The left and the right sides of the field are the sides of the equation. The "dragon box" is x, and you are trying to isolate it on one side. Galaxy images are 0, and can be removed at any time by tapping. Multiplication by 1 can similarly be done. Dice are used to represent numbers and can be added/subtracted, factored, divided/multiplied; other critters are used to represent other constants (a, b, c) and have the same done.

If you get far enough along, they substitute an 'x' icon for the dragon box, and letters for the constants.

However, the game doesn't explain why certain operations are allowed or disallowed -- each of the algebraic rules introduced is described as a "new magic power".

No offense, but clearly no good deed goes unpunished on Hacker News =)

Notwithstanding opinions on the merits or lack thereof of abstracting and visualizing concepts that may or may not help your syntactical understanding of algebra, I want to point out that this is A) a brilliant idea and B) a really cool implementation.

Kudos to the developer who spent the time to implement this, and to Brett for putting the idea forward. I hope this gets traction and is continued to be built upon.

It's great work. Constructive criticism can make it even better.

I know, and I don't mean to single you out, it just gets tiresome when the top-voted HN comment always seems to be critical. I felt this post in particular deserved better.

It was even more funny because while I was waiting for the comments to load I was thinking, "This is an awesome tool, but I bet the top comment is going to be some obscure reason not to like it."

So when I try to write + it writes = instead, not sure why.

It's tempting to call this 'programming' or 'code', in order to link those concepts. This is true. BUT

Consider the context of this kind of thing is different to the context in which we usually use code.

To explain:

It would be great if figures (e.g. in articles) came with code representing how they were calculated, such that they can be tweaked. In fact, I think actual programming source code should implement such a thing e.g. for constants. But this won't be part of the "real" code - the values will remain implemented as constants. Hence "code" like this should be still considered manual, a working tool, as opposed to the nature of common code - which is non-manual in the end...

There seems to be an issue where + is turned into = on my Swedish keyboard. So I can't type +.

edit: Can't seem to find any other character mapping issue. Perhaps this is due to some hardcoded mapping assuming an english keyboard?

I have the same problem with the + on my numpad.

Same as me on a British Mac keyboard number pad

I'm solving "2^... = 256" and it gives NaN. It works when the right hand side is smaller than 200 but fails when the right side is larger than 242.

Had some fun with this calculator toy. Scrubbing concept could work well in spreadsheet input values too. Maybe you could also scrub from a spreadsheet chart?

Text input doesn't seem to work for other than US keyboard layout. Does it read keycodes directly?

I've been meaning to make an online version of Calca.io / Soulver, which I love so much! If this is it, that'll be awesome.

Currently, I can't really use it much because the examples won't load (they flash, and then disappear)

Looks interesting, but doesn't seem to support my Swedish keyboard layout very well. Typing '+' inserts '=' instead. Pasting a '+' in works fine though.

Source here: https://github.com/osnr/cruncher

Getting an invalid cert for this via Chrome, anyone else having problems?

Just to point out: entering 1000^6 gives 999,999,999,999,999,900.

Reply on news.ycombinator.com