Comment by gpakosz | Hacker News Reader

Comment by gpakosz | Hacker News Reader

Hi folks,

A small note on what we don't support yet:

no cube roots, no nth roots either

square root's top line must be a single line (make it wide enough up front)

no matrices

no system of equations

no corrections, no scratch out (use top left undo/redo arrows)

leave enough space between integral/summation symbols and main expression for better accuracy

LaTeX output pleases MathJax as much as possible (thank you guys for your lovely rendering library)

Hope that helps, Thank you for trying it out.

PS: MathML seems to get little attention, why is it so?

--- LIST OF SUPPORTED SYMBOLS (encoded in UTF-8) ---

Letters

  a b c d e f g h i j k l m n o p q r s t u v w x y z
  A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Digits

  0 1 2 3 4 5 6 7 8 9

Maths symbols

  € $ £ ¥ ₩ ¢ ( ) < > [ ] { }  ! # % & ? @ / \ | ∥ © ∂∅ ∇ ∞ 
  ℂ ℕ ℚ ℝ ℤ
  + - ± × ÷ * ∘ · = ' , .  : ; _
  ← ↑ → ↓ ↔ ↕ ↖ ↗ ↘ ↙
  ⇐ ⇑ ⇒ ⇓ ⇔ ⇕
  ∀ ∃ ∄ ∈ ∉ ∋ ∌ ∩ ∪ ⊂ ⊃ ⊄ ⊅
  ∼ ≃ ≠ ≡ ≢ ≤ ≥ ≪ ≫ ∝ ∠
  ∏ ∑ ∫∮∧ √

Greek symbols

  Γ Δ Ω α β γ δ ε η θ λ ν π ρ σ τ φ χ ψ ω ϕ µ

International convention units (with cursive support)

  km hm dam dm cm mm µm
  ha
  hl dal dl cl ml µl
  kg hg dag dg cg mg µg
  ms µs
  GHz MHz kHz Hz

Other mathematical terms (with cursive support)

  sin cos tan sinh cosh tanh arcsin arccos arctan cot coth
  min max arg argmin argmax
  inf sup lim liminf limsup
  ln log
  dx dy dz dt

> PS: MathML seems to get little attention, why is it so?

MathML is hard to write and harder to read. The only time I pay any attention to MathML is when sites do not have TeX as an option. Others may feel differently, but this is my reason.

> MathML is hard to write and harder to read.

That's what tooling is for.

It looks like a great solution for including Math in html, only the browser support seems to suck. No Chrome or IE support without the use of a plugin.

> That's what tooling is for.

I disagree. Essentially requiring tools for a supposedly human-readable markup nullifies the whole human-readable aspect. Tools such as equation editors can be a great boon, but those can be had without giving up being able to read and write the markup directly.

I do think it's a shame that browser support for MathML is limited or missing. I'd much rather have direct, native MathML support than nothing. MathJax is a pretty nice way to address this for now, I think. Much better than requiring random site visitors to install a plugin.

I somewhat expected people would use it as an input to other math software

I think enough people are used to using LaTeX that MathML was never going to catch on, as its too verbose.

I was very impressed that it was able to interpret "min" with "x-2" underneath it, but if I wrote anything next to it the interpreter failed. I tried writing the "min" bit on one side and putting a very large parenthesis on the other side of the sheet, but it still didn't work.

Also, it handles \overline wonderfully -- impressively differentiating between an overline and fraction -- but it has trouble with the \underline. \Underbrace isn't working so well, either.

Yet another difficulty - "1 / 2" is turned into "y_2". The horizontal fraction delimiter doesn't seem to be recognized.

Very impressive!

It doesn't support '⊂'. It supports 'c' and 'C', but not '⊂'.

Edit: It's amazing work by the way, but distinguishing those definitely needs more training.

Writing '⊂' alone would likely be recognized as 'c' or 'C' because the decision is based on the context.

'A ⊂ B' is unkikely going to be recognized as 'A c B' or 'A C B'

is it a raster image recognition, or a vector/gesture recognition?

vector and gesture. This demo is a use case of our SDK. The SDK needs in input strokes (array of x/y) in the order. It returns a tree of symbols. We provides tools to transform into LaTex and MathML.

thanks, this is indeed very interesting. I gave a link at a Russian wearable computing community: http://ru-wearable.livejournal.com/107496.html

it's online recognition

we interpret the strokes (1 stroke == 1 sequence of 2D points)

in other words, it doesn't work on images; but we need input from a mouse, stylus, touchscreen, inkling, anoto pen, anything that outputs sequences of x,y coordinates

then it's rather a gesture recognition :)

it would be interesting to see it on a wearable computer, taking input from some gesture input devices: accelerometers or video input with markers

For some reason, the first thing I tried was circle-plus, which it thought was a theta followed by a bar or a one. D'oh.

Do you have an iOS SDK? What are the licensing terms? I'd love to integrate this into some of my math games.

It works splendidly with any algebra I throw at it, but it's not very much intended to handle expressions in logical calculus, is it?

Case in point: http://i.imgur.com/tnuxs.png

For a bug report like this, you might want to explain exactly what LaTeX you expect that to turn into.

$p,q \vdash p$ — sorry, I was pretty sure it was obvious to anyone, but I guess my handwriting wasn't that good.

Looks like it knows it's supposed to print "sin" and "cos" in non-italic text, but not the 'd' in \int x d x.

Extremely neat in any case.

I've almost never seen an upright 'd' in mathematical literature, though it seems to be somewhat more common in engineering texts. Ditto for the imaginary unit 'i' and the constant 'e'. Matter of taste.

All three are set upright in _The Princeton Companion to Mathematics_ (Gowers).

Hm. Now I wonder where I got the idea that of course it's supposed to be upright. Though I have been reading the Princeton Companion book mentioned in sibling.

I guess the problem is that the system drops the space between the expression and the dx, outputting the "xdx". Math text does have space before dx even if the letter is italic.

Cube roots and nth roots worked just fine for me.

You must be a real handwriting-wizard because I've tried for 10 minutes now without getting it to work! :(

Reply on news.ycombinator.com