[whatwg] Can <var> possibly work?

Ozob the Great ozob1337 at gmail.com
Sat Sep 20 14:19:52 PDT 2008

On Sat, Sep 20, 2008 at 4:13 AM, Kristof Zelechovski <giecrilj at stegny.2a.pl>

> In short, you should mark functions and constant parameters with VAR as
> well.
The spec says that <var> "represents a variable". To mark non-variables with
<var> is a misuse of markup; it's like surrounding computer documentation
with <code>.

> Scalar constants should be marked with I,
Set constants should be marked with B,
Set constants should not always be marked with B. Frequently they should be
italicized instead; see, e.g., Halmos, Naive Set Theory. And not all scalar
constants should be italicized; usually Greek letters are unitalicized
(e.g., the uppercase delta in the equation m = Δx / Δy).
Semantic markup does not suffice to determine presentation unless it's
extremely detailed; to give an example, if f is a function then it's usually
written in italics, as in <i>f</i>(<i>x</i>). But the trigonometric
functions sin, cos, tan, etc. are always written in roman type: sin

As I said before, MathML specifies about 120 different content identifiers.
In addition to what's on your list, it includes relations, functions,
vectors, matrices, and much more. It even distinguishes those pesky
roman-type trigonometric functions from ordinary functions. You might want
to take a look; the full list is at


> The VAR element does not indicate what the underlying text means
This is false. It "represents a variable". I concede that <var> may still be
appropriate for marking variables in computer code, but at present HTML
cannot accurately mark up mathematical content more complicated than x = y.
Making HTML that much richer would require duplicating a large chunk of
MathML, which is undesirable.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.whatwg.org/pipermail/whatwg-whatwg.org/attachments/20080920/5a822f36/attachment-0001.htm>

More information about the whatwg mailing list