You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
One simple and secure approach is to use the small mimetex (code mirrorred at https://github.com/icaoberg/mimetex) , which is a standalone renderrer, not using any other larger binaries like latex.
Another approach is the use the MediaWiki approach, which is also battle tested, of texvccheck/texvc , an OCaml program , which has a small checker program (texvccheck) that checks the AMS-Latex before sending to a second program that using latex to render it.
double click on the equation to be taken to the editor ;)
Reacted by André David, Yuchen Pei, Giannis Moutsinas, Zhitao Gong, Marcelo Lacerda, Unknown6656, Irene, Fangzhou Li, Sandeep Raju Prabhakar, Aziz Alto and 9 more
Reacted by Kevin, John D. Pope, Dror Atariah, Jun Mei, Jovan Sardinha and LM
It would be very easy to make a similar webapp to codecogs https://www.codecogs.com/latex/eqneditor.php. This would give the security isolation that GitHub needs by having it as a separate web based application.
It would be useful to attach persistant URL's to equations and use them for accessing the images and the editor rather than the long URL's codecogs uses.
Also codecogs does not support many packages. It does not support the 'semantic' and 'maths-symbols' that I need in order to make this useful and I am sure others will want other packages.
I taught that GitHub originally had MathJax on his pages, later on removed it to make sure everything was nice and responsive. Why would it be so difficult to have a "preview" button that enables mathjax on demand? Wikipedia seems to be doing math, yet maintain responsiveness, just fine!
Reacted by Kevin, Cheng-Yu Hsu, Aditya Gilra, Mirko Klukas, basblsolver, Michael Chirico, Xiao Han and Sachin Abeywardana
How about opt-in MathJax on a per-repository basis? General GitHub responsiveness would not be impaired, but simple math would be there for people who need it.
Reacted by Yuchen Pei, [object Object], Justus Sagemüller, unageek and Sachin Abeywardana
@allefeld anything really, but removing it cold-turkey creates a big problem.
Basically any mathematical software cannot use the wiki to document the software :(
This issue is definitely important, and it's really painful not to have support for equations.
Image based workarounds are clunky at best, hard to edit, and generally don't meet accessibility requirements, which are a legal mandate for all publicly funded universities in the US.
Individual browser plugin based workarounds don't help because they don't allow for someone else looking at the repo who doesn't have the plugin installed.
Native support is needed here.
If not this repo, then where? If it's code I can get to, I'm happy to work on it and do a PR if needed, but I'd want to know you all would merge once done before I dedicate the time. Let me know.
Reacted by Brendon O'Leary, Mitchell Paulus, Claudio Sousa, 0xf17, Hörmet Yiltiz, Torbjörn Rathsman, Ilya Baryshnikov, htdinh, Stefanos Carlström, Fritz Obermeyer and 3 more
I think taking a leaf out of wikipedia's book -
they do both - they have support for latex maths equations - and spit out an image with the 'alt' / text alternative markup allowing editors to update / edit / copy and paste.
Well here is a status report on that statement. There are four main browsers: Blink (WebKit fork for Chrome?), Edge, Gecko and WebKit. Gecko has complete MathML support and WebKit has experimental support. Safari (WebKit based) has it from 10.1. See also http://caniuse.com/#search=mathml. The Chrome and Edge teams have no plans to support it.
Then one can observe that XML is PITA and TeX is nice (from a user point of view, see https://groups.google.com/forum/#!topic/mozilla.dev.platform/96dZw1jXTvM). But the important thing here is that there should be a way of displaying formulas natively, without falling back to images. For Gecko, MathML is the way of doing it, while the other engines has no stable way.
Another point is that not all users are Pro(tm) users. If Chrome and Edge targets a consumer grade product, you may claim that math support is not needed. This is fine as long as there are browsers Pro(tm) users. Actually, one of the main reasons I use Firefox over Chrom* is that it supports math natively.
On 1 July 2017 at 20:00, Torbjörn Rathsman ***@***.***> wrote:
Well here is a status report on that statement. There are four main
browsers: Blink (WebKit fork for Chrome?), Edge, Gecko and WebKit. Gecko
has complete MathML support and WebKit has experimental support. Safari
(WebKit based) has it from 10.1. See also http://caniuse.com/#search=
mathml. The Chrome and Edge teams have no plans to support it.
Then one can observe that XML is PITA and TeX is nice (from a user point
of view, see https://groups.google.com/forum/#!topic/mozilla.dev.
platform/96dZw1jXTvM). But the important thing here is that there should
be a way of displaying formulas natively, without falling back to images.
For Gecko, MathML is the way of doing it, while the other engines has no
stable way.
Another point is that not all users are Pro(tm) users. If Chrome and Edge
targets a consumer grade product, you may claim that math support is not
needed. This is fine as long as there are browsers Pro(tm) users. Actually,
one of the main reasons I use Firefox over Chrom* is that it supports math
natively.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#897 (comment)>, or mute
the thread
<https://github.com/notifications/unsubscribe-auth/AAFOE5mmiaySAYdSxja1Mlg3DeDv5OQhks5sJpdegaJpZM4IrWxc>
.
--
Aaron Gray
Independent Open Source Software Engineer, Computer Language Researcher,
Information Theorist, and amateur computer scientist.
@milasudril , @AaronNGray : Shouldn't this discussion separate the markup users enter when editing and the result of rendering the markup? I mean, as far as the result is concerned, it could be MathML for WebKit and Gecko, and images for Edge and Blink. And as for the markup - that can be chosen half-independently and certainly does not depend on which browser you use.
But the important thing here is that there should be a way of displaying formulas natively, without falling back to images. For Gecko, MathML is the way of doing it, while the other engines has no stable way.
GitHub has good expertise in storage are and bad one in math apparently. Guys, it's not a way how to help users. Math formulas are tighten to the software more than you (probably) think.
Reacted by Royi, Fi3, Torbjörn Rathsman, Aleksei and Emile Mathieu
I'll repeat my last point — this repository is not for feature requests. It's a gem for delegating markup rendering.
Rendering math inline in e.g. Markdown is not a github-markup issue. You're welcome to reach out to the support team with further questions and suggestions about it, but please note that Markup does one specific thing — delegate rendering to gems — and rendering math inline in blobs happens elsewhere.
If you comment here, you don't reach the support team who can then direct your query appropriately, you reach me, and I'm not the person who decides what we implement in the GitHub product.
Activity
adavidzh commentedon Jun 29, 2016
Also #274 was asking to go in this direction. No chance to have a supported way to render math in github README files?
jayvdb commentedon Jul 1, 2016
I've been doing a little analysis on imsky/PocketEuler#44
One simple and secure approach is to use the small mimetex (code mirrorred at https://github.com/icaoberg/mimetex) , which is a standalone renderrer, not using any other larger binaries like
latex.https://packages.debian.org/search?suite=jessie&searchon=names&keywords=mimetex
https://admin.fedoraproject.org/pkgdb/package/rpms/mimetex/
No CVE since 20090713, which a batch of problems was found. See also https://www.debian.org/security/2009/dsa-1917.en.html
Another approach is the use the MediaWiki approach, which is also battle tested, of texvccheck/texvc , an OCaml program , which has a small checker program (texvccheck) that checks the AMS-Latex before sending to a second program that using
latexto render it.0 CVE : https://cve.mitre.org/cgi-bin/cvekey.cgi?keyword=texvc
jayvdb commentedon Jul 1, 2016
@bkeepers , you closed #274 saying "can't be resolved by a change to the code in this project".
So, which project(s) should this feature be added to? Each of the processors? Or a new project to do that math rendering only?
Most people are interested in maths in markdown , which is https://github.com/vmg/redcarpet , but they have rejected that feature request : vmg/redcarpet#313 and a smaller version of it vmg/redcarpet#395
RST and docutils supports maths, but this has been disabled #83
asciidoctor supports MathJax (asciidoctor/asciidoctor#492) , but without MathJax JS loaded by github markup it wont work. And there is https://github.com/asciidoctor/asciidoctor-latex
mediawiki has
<math>and wikicloth has supported it since 2011 (nricciar/wikicloth#19 / https://github.com/nricciar/wikicloth/blob/master/lib/wikicloth/extensions/math.rb) , but again it isnt enabled by github markup.Is there a plan? What is the plan? What do you need? How can we help?
jayvdb commentedon Jul 5, 2016
Note related request for MathML : #551
jayvdb commentedon Jul 10, 2016
Maybe a first step is for Github to treat typical math mode as non-markdown mode, similar to backticks (
..).Currently Github renders
Let $\text{S}_1(N) = \sum_{p=1}^N \text{E}(p)$with the
_denoting italics, as"Let $\text{S}1(N) = \sum{p=1}^N \text{E}(p)$"
The same thing happens with
*.AaronNGray commentedon Jul 10, 2016
@bkeepers I have found a stop gap solution a site https://www.codecogs.com/latex/eqneditor.php that generates embedded links with TeX editor that then embeds a gif in GitHub :-
double click on the equation to be taken to the editor ;)
adavidzh commentedon Jul 10, 2016
Hey @AaronNGray, that is pretty nifty!
adavidzh commentedon Jul 10, 2016
I'm wondering if the
codecogsworkaround of @AaronNGray could be packaged with github.io+MathJax?AaronNGray commentedon Jul 10, 2016
It would be very easy to make a similar webapp to codecogs https://www.codecogs.com/latex/eqneditor.php. This would give the security isolation that GitHub needs by having it as a separate web based application.
It would be useful to attach persistant URL's to equations and use them for accessing the images and the editor rather than the long URL's codecogs uses.
Also codecogs does not support many packages. It does not support the 'semantic' and 'maths-symbols' that I need in order to make this useful and I am sure others will want other packages.
https://www.ctan.org/tex-archive/macros/latex/contrib/semantic
https://www.ctan.org/pkg/maths-symbols
The following should come out as an inference rule :-
and codecogs does not support \usepackage{}
AaronNGray commentedon Jul 10, 2016
I am putting together a Node.js solution for this.
taiya commentedon Jul 31, 2016
I taught that GitHub originally had MathJax on his pages, later on removed it to make sure everything was nice and responsive. Why would it be so difficult to have a "preview" button that enables mathjax on demand? Wikipedia seems to be doing math, yet maintain responsiveness, just fine!
FlorianFranzen commentedon Aug 1, 2016
+1
allefeld commentedon Aug 7, 2016
How about opt-in MathJax on a per-repository basis? General GitHub responsiveness would not be impaired, but simple math would be there for people who need it.
taiya commentedon Aug 7, 2016
@allefeld anything really, but removing it cold-turkey creates a big problem.
Basically any mathematical software cannot use the wiki to document the software :(
AaronNGray commentedon Aug 7, 2016
I will see if I can get a third party TeX![]()
based solution in place ASAP. I got distracted trying to get Angular2 working in order to do the UI.
48 remaining items
nihonjinrxs commentedon Jun 15, 2017
This issue is definitely important, and it's really painful not to have support for equations.
Image based workarounds are clunky at best, hard to edit, and generally don't meet accessibility requirements, which are a legal mandate for all publicly funded universities in the US.
Individual browser plugin based workarounds don't help because they don't allow for someone else looking at the repo who doesn't have the plugin installed.
Native support is needed here.
If not this repo, then where? If it's code I can get to, I'm happy to work on it and do a PR if needed, but I'd want to know you all would merge once done before I dedicate the time. Let me know.
johndpope commentedon Jun 29, 2017
Well - some good news - thanks to @goessner - and his markdown math plugin for Visual studio code - you can write / read maths equations quite elegantly by the look of it
https://marketplace.visualstudio.com/items?itemName=goessner.mdmath
Dear Github - please roll out Markdown maths in gists.
https://github.com/goessner/mdmath
johndpope commentedon Jun 30, 2017
I think taking a leaf out of wikipedia's book -
they do both - they have support for latex maths equations - and spit out an image with the 'alt' / text alternative markup allowing editors to update / edit / copy and paste.
for anyone using Chrome extension - you may want to check out
Alt Text Tester (where you can double click to copy latex ) - https://chrome.google.com/webstore/detail/alt-text-tester/koldhcllpbdfcdpfpbldbicbgddglodk
there is also this chrome plugin to spit out images of maths equations https://chrome.google.com/webstore/detail/equatio-math-made-digital/hjngolefdpdnooamgdldlkjgmdcmcjnc
milasudril commentedon Jul 1, 2017
However, native mml should be supported as well. Wikipedia doesn't.
AaronNGray commentedon Jul 1, 2017
milasudril commentedon Jul 1, 2017
Well here is a status report on that statement. There are four main browsers: Blink (WebKit fork for Chrome?), Edge, Gecko and WebKit. Gecko has complete MathML support and WebKit has experimental support. Safari (WebKit based) has it from 10.1. See also http://caniuse.com/#search=mathml. The Chrome and Edge teams have no plans to support it.
Then one can observe that XML is PITA and TeX is nice (from a user point of view, see https://groups.google.com/forum/#!topic/mozilla.dev.platform/96dZw1jXTvM). But the important thing here is that there should be a way of displaying formulas natively, without falling back to images. For Gecko, MathML is the way of doing it, while the other engines has no stable way.
Another point is that not all users are Pro(tm) users. If Chrome and Edge targets a consumer grade product, you may claim that math support is not needed. This is fine as long as there are browsers Pro(tm) users. Actually, one of the main reasons I use Firefox over Chrom* is that it supports math natively.
AaronNGray commentedon Jul 1, 2017
eyalroz commentedon Jul 1, 2017
@milasudril , @AaronNGray : Shouldn't this discussion separate the markup users enter when editing and the result of rendering the markup? I mean, as far as the result is concerned, it could be MathML for WebKit and Gecko, and images for Edge and Blink. And as for the markup - that can be chosen half-independently and certainly does not depend on which browser you use.
milasudril commentedon Jul 2, 2017
@eyalroz That was the implied meaning of
andy-shev commentedon Jul 10, 2017
GitHub has good expertise in storage are and bad one in math apparently. Guys, it's not a way how to help users. Math formulas are tighten to the software more than you (probably) think.
RoyiAvital commentedon Jul 11, 2017
Any chance to make GitHub reconsider this?
Today many of the codes in GitHub deals with Data Analysis and Math documentation is essential for that.
Please reconsider.
FlorianFranzen commentedon Jul 11, 2017
@kivikakk Since you have not replied by now, I assuming your decision is final. Gitlab it is.
sachinruk commentedon Sep 1, 2017
ping. Any update on this? I dont think anyone has given up on wanting math equations on their README's.
kivikakk commentedon Sep 1, 2017
I'll repeat my last point — this repository is not for feature requests. It's a gem for delegating markup rendering.
If you comment here, you don't reach the support team who can then direct your query appropriately, you reach me, and I'm not the person who decides what we implement in the GitHub product.