Inverse graphing calculator | Hacker News Reader

Inverse graphing calculator

xamuel.com

38 pointsanigbrowl16 comments

How about a copy-able version of the equations so that we can test them to see if they actually work?

Seconding this. Plus then you can be really nerdy and say "yo, graph this" and send the equation to them and they type it into something and see the message.

Usecase: if this had come out like 2 weeks ago with copy-able equations, I probably would have asked someone to prom with it.

The URL of the equation image contains the equation in TeX.

Copy image location of the img gives you the equation in TeX.

Then you can use ToExpression[...] in Mathematica to try and graph it.

I haven't tried it myself, but I assume that should work.

http://reference.wolfram.com/mathematica/tutorial/Generating...

Hm, Wolfram|Alpha doesn't like that function, and I don't have access to a real instance of Mathematica...

Not so good.

First minor hurdle: Mma treats backslashes in strings as escapes. But even after doubling them all, I still get "Could not parse [BLAH] as Mathematica input". A bit of experimenting suggests that Mma doesn't like \cdot, so I removed those. I took out the explicit linebreaks at the same time.

Now Mma is able to parse the TeX. So far as I can tell from a quick glance, it's understood it OK. Unfortunately, I've had no success in getting it to plot it. RegionPlot[...==0...] produces nothing. ContourPlot[...==0,...] produces nothing. Without the ==0, unsurprisingly, no contour Mma chooses to draw looks at all appropriate. The trouble is that (1) the function is varying so damn fast in the relevant region and (2) those zero contours are all "valley-bottoms", and there's not much way for Mma to know they're there at all. Even giving unreasonably large values for WorkingPrecision,PlotPoints and MaxRecursion doesn't help, unless you consider running the Mathematica kernel out of memory "helping".

This is clever. But surely, 'l' and 'o' should be simpler.

Someone make a self printing equation.

It would totally make my day.

Here you go: http://en.wikipedia.org/wiki/Tuppers_self-referential_formul...

I hope your day is made now :)

Nope. That's an arbitrary-thing-printing equation, controlled by a parameter that you have to set to a Very Large Number to make it display anything interesting. The Very Large Number is not part of what the formula displays. To get a genuinely self-displaying formula you'd need to throw in some Goedel/Quine tricksiness.

Very neat.

I'm guessing you have a function for each letter and you add a shift depending on where it is in the phrase? Or are you creating the formulas on the fly?

If its being created dynamically, you could probably find a cool way to do this for arbitrarily drawn pictures.

Why does every "string" end in (y^2 - 6y +8 + sqrt(y^4 - 12y^3 + 52y^2 - 96y + 64))^2 ?!?

The empty string does not.

http://www.xamuel.com/inverse-graphing-calculator.php?phrase...

But, I have no idea. Perhaps it is a bug wherein a CRLF is being added to the internal string? However, I highly doubt that.

It's a rather complicated way to ensure that all points have y coordinate between 2 and 4, thus cutting all lines that would extend out of the "writable" two squares.

That expression equals zero if and only if 2<=y<=4.

(edit) If you try "L", the formula contains a product member (x-2) that would produce a vertical line whose points all have x=2, containing the vertical segment of "L". The final string excludes that.

how about an "inverse" graphing calc that takes an arbitrary line drawing and gives me a "nice-looking" equation and not just the vectorized representation?

Great now I can just generate the equations for "suck my cock" and put them on a t-shirt. And then I will be insulting everyone I meet without them even knowing.

Reply on news.ycombinator.com