W3C home > Mailing lists > Public > www-math@w3.org > April 2004

Re: Combing markup

From: Carole Snyder <carolersnyder@hotmail.com>
Date: Fri, 16 Apr 2004 06:36:56 -0700
To: <www-math@w3.org>
Message-ID: <Sea1-DAV47cU6xfPTei00010193@hotmail.com>
Oh dear.  I realized my lengthy message did not wrap, at least not in the archives.
Not very user friendly or Web savvy.  I've reposted the message, for those of you that received this in a readable format the first time, I'm sorry for the spam.  (Strange that this happened.  I wrote the message in MS Word and pasted it in Outlook Express.)

Original Message:

I’m new to MathML, and want to check my understanding of the specification.  I’ve been 

researching MathML in preparation to use it to allow our math editing software to 

communicate with other applications.  It seems that there are various ways that content and

 presentation markup can be included in the same document, and applications are free to 

produce and read combined markup in a way that makes the most sense for the purpose of the 

application.  Each method of combining markup is useful for different reasons, and 

application developers should decide which methods are most useful for their application.  I 

have gone through several methods of combining markup and indicated, to the best of my 

knowledge, where and why each method might be used.  I’m hoping that someone will read and 

respond to my notes, and provide feedback, indicating where I am wrong or missing important 

points concerning the purpose of the combination method.


* Mixed Markup, as described in section 5.2 of the specification, is useful when an application is 

primarily either for presentation or computation.  One would choose one type of markup as the

 “primary” markup, and embed the other type of markup as “hints”.  So if an application renders 

MathML that is meant primarily for viewing, the application would use mostly presentation 

markup, and include content markup to provide any meaning that is not immediately obvious, or 

the default semantic for the presentation.  For example, “f(x)” can be thought as “F times quantity 

x,” instead of “F of x,” which is a more common usage of the notation.  Including the content 

markup in the presentation markup clarifies the meaning.  (Computation-oriented applications will 

want to include presentation markup when a particular expression should be presented a certain 

way.  For example, should 1/x be written as a fraction, or x to the power of –1?)


* Top-level parallel markup, section 5.3.1, has two parallel branches. One is presentation markup, 

and the other is content markup.  This allows an application to present the expression in the 

desired form, as well as communicate the semantics of the expression.  The limitation of top-level 

parallel markup is that the correspondence between the content and presentation markup is only 

for the entire expression.  If an application wants to manipulate (copy, edit, etc.) sub expressions, 

top-level parallel markup does not indicate how a sub expression in one branch corresponds to 

the other branch.  Each sub expression can have a presentation branch and a content branch, 

but this can cause the MathML document to become very large.


* Parallel markup via id and xref allows for the MathML document to have one branch for each 

type of markup, and still allows the application to deal with sub expressions.  Both the content 

markup and the presentation markup are present, and an application can match a sub expression 

in one branch to its counterpart in the other branch.


* From section 7.2.1: “developers are given wide latitude in interpreting what kind of MathML 

implementation is meaningful for their own particular application.”  This means that applications 

can choose how to respond to various ways MathML is authored.  For example, a content-

oriented application interpret presentation markup that is embedded in content markup, as well as 

display the expression correct when a document uses top-level parallel markup, but the 

application does not manipulate sub expressions when id and xref attributes are present.  Also, 

the same application can produce a MathML document that uses only content MathML.


* If the primary goal is to communicate with other applications that support MathML, one needs to 

consider which types of applications one wishes to communicate with, so the MathML rendered is 

the one that’s best suited for that type. Parallel Markup will probably be the most flexible choice, 

because it allows an application to parse either the content markup or the presentation markup, 

or use both simultaneously.  But it’s not necessarily always the best choice (it might be overkill for 

one’s purposes.)


I look forward to any feedback you have.  Thank you for your time.



Carole Snyder

Henter Math
Received on Friday, 16 April 2004 09:37:36 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:27:35 UTC