Forgot your password?
typodupeerror
Firefox Chrome Math Software

Firefox Is the First Browser To Pass the MathML Acid2 Test 134

Posted by timothy
from the assembling-data-please-wait dept.
An anonymous reader writes "Frédéric Wang, an engineer at the MathJax project, reports that the latest nightly build of Firefox now passes the MathML Acid2 test. Screenshots in his post show a comparison with the latest nightly Chrome Canary, and it's not pretty. He writes 'Google developers forked Webkit and decided to remove from Blink all the code (including MathML) on which they don't plan to work in the short term.'"
This discussion has been archived. No new comments can be posted.

Firefox Is the First Browser To Pass the MathML Acid2 Test

Comments Filter:
  • by corychristison (951993) on Friday May 03, 2013 @10:30PM (#43626609)

    My many web design/development clients would disagree with you. I don't even want to recall the times I've had to tell them No for blinking things.

    Unfortunately they think blinking == attention getting, whereas we think blinking == f*cking irritating.

  • by Anonymous Coward on Friday May 03, 2013 @11:23PM (#43626859)

    Personally I browse pages that use MathML every single day (literally), opengl.org ref pages, have MathML scattered throughout - I've never bothered installing chrome on a workstation once due to lack of MathML support, and don't ever plan on doing so until it can (among many other missing features).

  • by dgatwood (11270) on Friday May 03, 2013 @11:24PM (#43626865) Journal

    Chrome stable passed it just fine.

    Are you looking at the same MathML Acid tests that I am? These are not the same tests as the original Acid and Acid 2 tests.

    Chrome Stable, at least on OS X, fails MathML Acid 1 miserably, as does Chrome Canary. All the fractions are shown as the left part followed by a space followed by the right part instead of as a fraction. Safari on OS X is basically correct for MathML Acid 1 (albeit with the ugliest parentheses I've ever seen), but if you reload the page, the curly braces disappear, as do the comma inside the curly braces and the plus sign.

    And the MathML Acid 2 test looks like a freaking Picasso on all three browsers.

  • by tyrione (134248) on Saturday May 04, 2013 @12:20AM (#43627077) Homepage

    The reason to improve MathML support isn't browsers. It's eBook readers that share the underlying rendering engines. Lots of textbook publishers want to use MathML, but without robust, reliable, visually appealing support, everybody has to do awful hacks with inline images or (hopefully) SVG instead.

    Spot on. EPub 3.x with MathML 3.0 is vital to work in digital publishing for iOS, Android, etc.

  • by TheRaven64 (641858) on Saturday May 04, 2013 @04:16AM (#43627763) Journal

    MathML attempts to separate the content and presentation. This is fine if you have a tool that properly supports both (I've never used one, maybe Mathematica or similar does?), but it sucks for most editors. The idea is that you can have a single format that describes both how to lay out equations and their semantics. In practice, pretty much everyone who generates MathML does it from the TeX equivalent and so only ever gets the presentation form. The other advantage of MathML is that each individual element is exposed via the DOM, so it's easy to manipulate equations from JavaScript, although I don't think I've ever seen that done either.

    Part of the problem with a format that is basically impossible for humans to write is that it also ends up being difficult to produce tools that can write and display it, which is why it's taken 10 years or so for MathML to get even a token amount of support in mainstream browsers...

  • by maxwell demon (590494) on Saturday May 04, 2013 @10:45AM (#43629091) Journal

    Just a note—sin^2(x) cannot be sin(sin(x)) because that is a datatype error.

    You just showed that you don't know enough mathematics.

    The input to the sine function is not an angle, it is a real or complex number. If real, this number often (but not always!) describes some angle. If complex, it obviously won't describe an angle.

    The sine function is defined as

    sin x = (exp(i x) - exp(-i x)) / (2i)

    where i is the imaginary number, and exp(x) is defined by the series

    exp x = 1 + x + x^2/2 + ... + x^n/n! + ...

    Note that, since the convergence radius of the exponential series is infinite, and the sine is just a linear combination of exponentials, the sine is defined on all complex numbers. Since it is complex-valued, sin sin x is indeed well defined for all complex numbers x.

    Moreover, if you restrict the sine to real numbers (that is, only accept real numbers), you still have a well defined sin sin x, because the real sine function is also real-valued (more exactly, its values are restricted to the interval [-1,1]).

    Also, the output is in general not a rational number (the only thing you could have meant with "ratio" that makes sense in this context).

"Irrationality is the square root of all evil" -- Douglas Hofstadter

Working...