Visual line breaks or alignment of equations

[TimArborealis]

I have noticed that the vendors often use <mtable> for visual line breaks or alignment of equations. While the visual result is often fine, it leads to excessive verbosity when read by TTS. Example:

is read out "Two lines, line one, the sum [...]". This can quickly become annoying when there are many big equations in a book.

Are there better ways to deal with visual line breaks and/or alignment? If we decide to advice against this, it would be good if the Kvalidator could check for it.

[MartheGjelstad]

The following is mentioned in issue #75 , but it make more sense to write it here.

Rules in MathCAT for speaking rows:

For ClearSpeak you can change the rule with name "ClearSpeak-default-multiline" and tag "[mtr, mlabledtr]". If you for example add "parent::m:system-of-equations" inside "not" in the first if statement after "replace", the word row and the row number will not be spoken for systems of equation and equation solving notation.

Another solution is to change the preferences in the file "prefs.yaml" for the line "MultiLineLabel". Change from "auto" to "none". Then only the number of rows will be spoken in the beginning. This will happen for both systems of equations and piecewise functions. See the test file ClearSpeak/multiline. There the different preferences are used in the tests.

Also, users can change this preference, even thought it's not part of the visible settings in NVDA.

I can investigate things further if desired.

---

For expressions where there are line breaks beacause of the page width, I don't think there should be line breaks in the markup, and there should be no <mtable>. However, for equation solving notation, where the equal signs are aligned, I think <mtable>s should be used.

[TimArborealis]

Thanks for pointing me to prefs.yaml, I hadn't tinkered with it before. Looks like the solutions to several context misinterpretations in our book production could lie here. I think it would be great if we could have access to these settings in WebArch so that we could change them for each TTS generation if need be. I'll make a note of that for future discussions with TexTalk.

However, when I tried setting MultilineLabel to None in my local MathCAT install, there was no change in what was spoken. I might be doing somthing wrong though and would much appreciate if you could demonstrate this, perhaps at the next meeting.

Regarding the ClearSpeak rule change, I think it might be a perilous road to create our own "custom" version of MathCAT since it could make the task of keeping it up to date more difficult. So unless we are sure that most users would benefit from the change (and if we are, we should suggest a change of the English rules), perhaps we should leave this alone.

Regarding your last point, I agree and I think we should put this in the guidelines.

[MartheGjelstad]

I can demonstrate in the next meeting :)

I agree about not changing the ClearSpeak rules. Since it's possible to change the preferences on a user level, I don't think it's necessary to change the rules in the yaml files.

[MartheGjelstad]

Test out the use of horizontal scrolling. Look at https://adrianroselli.com/2017/11/a-responsive-accessible-table.html

[samimaattaCelia]

@AndersEkl, I made a demo on my site: Testing math overflow. I have not tested this in a reading system.

Some notes about using it in the browser (Firefox):

• Mouse accessibility? It can be scrolled with the mouse.
• Keyboard accessibility? Can be tabbed into and moved with arrow keys.
• Screen reader accessibility? NVDA+MathCAT can access the math, but does not scroll with the TTS. It probably won't follow the highlight in a book with a media overlay.

Code

It uses a CSS style:

.math-block-wrapper {
overflow-x: auto;
overflow-y: hidden;
}

Here is the markup (which I just copied from somewhere, so it's not by the guidelines):

<div class="math-block-wrapper">
      <math display="block" xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle P(x)=x^{n}+b_{1}x^{n-1}+\cdots +b_{n-1}x+b_{n}=(x-x_{1})\cdots (x-x_{n}).}">
  <semantics>
    <mrow class="MJX-TeXAtom-ORD">
      <mstyle displaystyle="true" scriptlevel="0">
        <mi>P</mi>
        <mo stretchy="false">(</mo>
        <mi>x</mi>
        <mo stretchy="false">)</mo>
        <mo>=</mo>
        <msup>
          <mi>x</mi>
          <mrow class="MJX-TeXAtom-ORD">
            <mi>n</mi>
          </mrow>
        </msup>
        <mo>+</mo>
        <msub>
          <mi>b</mi>
          <mrow class="MJX-TeXAtom-ORD">
            <mn>1</mn>
          </mrow>
        </msub>
        <msup>
          <mi>x</mi>
          <mrow class="MJX-TeXAtom-ORD">
            <mi>n</mi>
            <mo>−<!-- − --></mo>
            <mn>1</mn>
          </mrow>
        </msup>
        <mo>+</mo>
        <mo>⋯<!-- ⋯ --></mo>
        <mo>+</mo>
        <msub>
          <mi>b</mi>
          <mrow class="MJX-TeXAtom-ORD">
            <mi>n</mi>
            <mo>−<!-- − --></mo>
            <mn>1</mn>
          </mrow>
        </msub>
        <mi>x</mi>
        <mo>+</mo>
        <msub>
          <mi>b</mi>
          <mrow class="MJX-TeXAtom-ORD">
            <mi>n</mi>
          </mrow>
        </msub>
        <mo>=</mo>
        <mo stretchy="false">(</mo>
        <mi>x</mi>
        <mo>−<!-- − --></mo>
        <msub>
          <mi>x</mi>
          <mrow class="MJX-TeXAtom-ORD">
            <mn>1</mn>
          </mrow>
        </msub>
        <mo stretchy="false">)</mo>
        <mo>⋯<!-- ⋯ --></mo>
        <mo stretchy="false">(</mo>
        <mi>x</mi>
        <mo>−<!-- − --></mo>
        <msub>
          <mi>x</mi>
          <mrow class="MJX-TeXAtom-ORD">
            <mi>n</mi>
          </mrow>
        </msub>
        <mo stretchy="false">)</mo>
        <mo>.</mo>
      </mstyle>
    </mrow>
    <annotation encoding="application/x-tex">{\displaystyle P(x)=x^{n}+b_{1}x^{n-1}+\cdots +b_{n-1}x+b_{n}=(x-x_{1})\cdots (x-x_{n}).}</annotation>
  </semantics>
</math></div>

[MartheGjelstad]

Related to issue #61

[MartheGjelstad]

I think this image is a good example of a case where an <mtable> should not be used for line break.

[MartheGjelstad]

Documenting my testing:

Used the suggested CSS style and <div class="math-block-wrapper"> around the block math element.

• Edge: Looks okay
• Firefox: The expression is separated on two lines
• Chrome: Looks okay
• Tibi: Changed CSS: The expression is bigger and you can scroll. Not changed CSS: You can scroll. Changed CSS and produced TTS: No automatic scrolling when the expression is read.
• EasyReader: No changes
• Validates in Webarch

Question

In previous meetings, we have discussed when to markup expressions on multiple lines. For example, one suggestion is that an expression should be marked up on multiple lines if the equal signs are alligned under each other. We have then talked about that it could be difficult for the suppliers to decide when to split an expression in mutliple lines and when not to.

If we have to use <div class="math-block-wrapper"> to scroll inside an expression, then we have the same issue that the suppliers have to decide when to use this markup and when not to.

[AndersEkl]

@MartheGjelstad did you run this in ACE?