Dear all,

I've got a quick question about the use of subscripts in Content MathML which I hope someone can shed some light on.

(a) Background:

Basically, this problem arose with the study of sums; it is often useful to describe a summation using a single algebraic letter with a subscript denoting the exact term. For example:

Sum ( a_i ) = a_1 + a_2 + a_3 + ... + a_n
i = 1

where the a_i notation means "a subscript i". There are lots of instances where this notation is used, and I'm sure everyone has come across it before; one elementary example is in series (for example, see

(b) Problem

The MathML spec. suggests than any need for subscripts within Content markup should be made using Presentation embedded within Content. For example, one way to describe a_i would be:


The problem with this is that the identifiers 'a' and 'i' are presentation-only, and hence could not be operated on by our algebraic processor engine (which uses a Content-only DOM).

(c) Solution

What we would like to do is the following:


Here we're mixing the two types of mark-up by embedding Content within Presentation within Content. However, when the processing engine evaluates the sum, it can easily replace all occurances of <ci>i</ci> with <cn>1</cn>, <cn>2</cn> etc... which would give the required terms (a_1, a_2 etc...) in the sum.

My question is really whether this is VALID mark-up? I believe it probably is [DTD definition p.301 MathML 2.0 spec., entities %PresExpression and %ContInPres] but I want to make sure with the experts that I am correct before proceeding.

Another question: in order to render mixed mark-up in a user-agent, would an XSLT transform need to be applied on the content elements?

Thanks in advance,
Charles Lyons. Anti-Virus & Anti-Spam Web Mail thats hotter than hot