W3C home > Mailing lists > Public > public-mathonwebpages@w3.org > December 2016

Re: plain text math notation

From: Han <laughinghan@gmail.com>
Date: Thu, 1 Dec 2016 03:00:18 -0800
Message-ID: <CACE=nd2pbJcAu2WWWT4D0rQVhaOC8ANm7qnO+o4ByhV6BSwEjg@mail.gmail.com>
To: Kevin Cheung <KEVINCHEUNG@cunet.carleton.ca>
Cc: Paul Topping <pault@dessci.com>, Jos de Jong <wjosdejong@gmail.com>, Robin Berjon <robin@berjon.com>, Peter Krautzberger <peter.krautzberger@mathjax.org>, "public-mathonw." <public-mathonwebpages@w3.org>
>
> It will be nice if there is a math notation language that forces math
> writers to mean exactly what they write.  (Contrasting to what
> mathematicians sometimes do; they use the same LaTex symbols to mean
> different things in different contexts.)  With CAS, since the goal is to
> compute, you have no choice but to type what you mean.
>
Do you consider CAS's to be examples of such languages? What about Content
MathML and OpenMath, or is this a third category, distinct from the Content
vs Presentation MathML categories that (I think) quite cleanly divide all
the notations mentioned so far?

(To me, it's clear that:

Presentation MathML-likes include: LaTeX, troff/eqn
<https://en.wikipedia.org/wiki/Eqn>, the LibreOffice syntax which looks
very similar, possibly identical to troff/eqn, AsciiMath, UnicodeMath,
MathSON (I'm the one working on MathSON, btw)

Content MathML-likes include: OpenMath (did they merge or something?),
Mathematica, Matlab, Octave, mathjs, Maple, Excel, SymPy, algebra.js

Note that except for OpenMath and Content MathML itself, all the content
formats are directly tied to a CAS and/or evaluation engine; and except for
Presentation MathML itself, all the presentation formats are directly tied
to a rendering engine.

Also note that while all the content formats are meant for machine
evaluation or manipulation and are therefore tree-structured and quite
amenable to machine manipulation, the only presentation format that is
tree-structured and hence amenable to machine manipulation is MathML, if we
ignore MathSON which basically doesn't exist yet (zero implementations);
which is kind of exactly why I'm working on MathSON.)

Han

> ------------------------------
> *From:* Paul Topping <pault@dessci.com>
> *Sent:* Thursday, September 22, 2016 3:17:56 PM
> *To:* Jos de Jong; Robin Berjon
> *Cc:* Peter Krautzberger; public-mathonw.
> *Subject:* RE: plain text math notation
>
>
> The goal of a single math language that describes standard math notation
> AND works well for computation and manipulation is a worthwhile and, IMHO,
> achievable goal. However, starting at the syntax is not going to work.
> Sure, it is fine to express preferences for one language or another but the
> place to start is the model that the language expresses. By model, I mean
> abstract data structure and its dynamic semantics.
>
>
>
> MathML suffered from the too-many-cooks problem as well as not having a
> clear set of goals, IMHO. It also suffered from its association with XML
> but that’s a whole other story.
>
>
>
> I believe starting with the model is the only road to success but it is
> perhaps hard to convince people of that. Another advantage that might be
> more compelling is that, with well-designed model in hand, it should be
> possible to develop multiple syntaxes for expressing it, each useful in a
> different scenario: an XML syntax for the XML document world and, perhaps,
> for embedding in HTML, a JSON syntax for programmatic generation,
> manipulation, and consumption, a “plain-text math notation” for authors to
> type directly, etc.
>
>
>
> BTW, I am not sure “plain text math notation” is well-defined. In a sense,
> all computer languages are plain text and, at the same time, all have a
> syntax. I’m guessing what is meant by this is a notation that humans feel
> comfortable typing. Computers like regular syntaxes and are not bothered by
> extra characters such as angle brackets. Humans like something that doesn’t
> require too many keystrokes, hard to reach keys on the keyboard, doesn’t
> get scrambled in email, and is easy for humans to read. If this is what is
> meant, perhaps “human math language” might work. Just a thought.
>
>
>
> Paul Topping
>
>
>
> Design Science, Inc.
>
> "How Science Communicates"
>
> Makers of MathType, MathFlow, MathPlayer, Equation Editor
>
> http://www.dessci.com
>
>
>
> *From:* Jos de Jong [mailto:wjosdejong@gmail.com]
> *Sent:* Thursday, September 22, 2016 11:52 AM
> *To:* Robin Berjon <robin@berjon.com>
> *Cc:* Peter Krautzberger <peter.krautzberger@mathjax.org>;
> public-mathonw. <public-mathonwebpages@w3.org>
> *Subject:* Re: plain text math notation
>
>
>
> @Robin can you elaborate a bit on your idea? Do you mean that it would be
> trivial to translate a math formula into plain JavaScript which then can be
> evaluated, or to SVG for rendering? Or both?... Do you have concrete use
> cases in mind?
>
>
>
> Jos
>
>
>
> On Thu, Sep 22, 2016 at 3:53 PM, Robin Berjon <robin@berjon.com> wrote:
>
> On 22/09/2016 02:54 , Peter Krautzberger wrote:
> > Jos has written up some notes on plain text math notations at [1].
>
> This is IMHO a very promising area. I wonder if it may be interesting to
> consider making it not too distant from the syntax of (modern) JS so
> that it could easily be implemented as a BabelJS[0] plugin (like JSX[1]).
>
> [0] http://babeljs.io/
> [1] https://facebook.github.io/react/docs/jsx-in-depth.html
>
> --
> • Robin Berjon - http://berjon.com/ - @robinberjon
> • http://science.ai/ — intelligent science publishing
> •
>
>
>
Received on Thursday, 1 December 2016 11:01:09 UTC

This archive was generated by hypermail 2.3.1 : Thursday, 1 December 2016 11:01:09 UTC