Grants:IdeaLab/Transform LaTex formulas in spoken words and make it readable with text to speech programs.

From Meta, a Wikimedia project coordination wiki
Transform LaTex formulas in spoken words and make it readable with text to speech programs.
Create a software that transforms LaTex's code in spoken words in order to make mathematics and the listening of formulas accessible to blind and dyslexic people who normally use speech programs to access to the knowledge.
idea creator
MassimoGiordano
researcher
Akiva.avraham
this project needs...
volunteer
community organizer
join
endorse
created on10:09, 1 October 2014 (UTC)


Project idea[edit]

What is the problem you're trying to solve?[edit]

Mathematics is still not completely accessible to blind and dyslexic people that "to read" they have to use “text to speech” programs that convert text into listenable spoken words. Doing that they are able to understand through listening. All the mathematical knowledge contains many formulas that “text to speech" programs cannot interpret. This generates huge gaps in mathematical concepts' understanding. In general, this kind of programs start to pronounce incomprehensible words instead of the formulas' contents, causing concentration losses for the reader.

What is your solution?[edit]

The main idea to solve this problem is to create a software that is able to transform the latex code in spoken word. To do this we can start with a library for each languages that users can expand. This library provides the exact sequence of spoken words associated with a particular structure in latex.

We make an example:

We can have a formula like this:
where the latex's code is: \int_{a}^{b} \frac{1}{\sqrt{\rho}} d\rho
The output that we want to have is: "The integral from a to b of one over the square root of ro in de ro. "

A basic example of library could be:

Latex's syntax output related
\int_{①}^{②} ③ d④ Integral from ① to ② of ③ in de ④
\frac{①}{②} ① divided by ②
\sqrt{①} square root of ①
\sqrt[②]{①} ②-st root of ①
\rho ro

The idea is that using this library the compiler know how to build the sentences placing the words in the correct order. Recursively it converts all the code:

    The integral from a to b of \frac{1}{\sqrt{\rho}}  in de ro
The integral from a to b of one divided by \sqrt{\rho} in de ro
The integral from a to b of one divided by square root of ro in de ro.

It is also possible add some statement in order to describe more different cases:

Latex's syntax conditions output related
\frac{①}{②} if ① == 1 one over ②
if ① != 1 ① divided by ②

Words and sentences that are generated must be readable by “text to speech" programs, for this reason the element alt in the HTML code have to contain words instead of Latex code as it is now.

  <img class=" " alt=" \LaTeX " src="  ">  <img class=" " alt="  words  " src="  ">

This would remove a useful feature of wikipedia pages. In fact, by simply copying the image and pasting it into a text editor, it is possible get the latex code of a formula to implement own work. An idea to fix this is create a new stander to compile the source of the page. As for the mobile version we use "en.m.wikipedia.org" we could use "en.v.wikipedia.org" for the listenable version. Where v means voice.

Progress[edit]

Get Involved[edit]

Participants[edit]

  • Developer Currently developing a library in Qt. Akiva.avraham (talk) 11:38, 26 April 2016 (UTC)
  • Researcher Doing the research for it as well. Akiva.avraham (talk) 11:42, 26 April 2016 (UTC)
  • Developer Respected Madam,
This information is useful for me and i am doing this project in Python.i need more information for doing this project.

Kindly forward any papers or material related to this project, My mail id:suganya-cse@saranathan.ac.in Suganyahariharan (talk) 04:36, 28 December 2017 (UTC)

Endorsements[edit]

  • Great idea Pengo (talk) 07:51, 10 March 2015 (UTC)
  • This idea would make mathematics a real lot more accessible to the blind and dyslexic. The positive effects of this are obvious. Mathmensch (talk) 09:48, 18 March 2016 (UTC)