Is invisibleTimes mandatory? (Was Re: Profiling ...)

<juanrgonzaleza@canonicalscience.com> writes:

> In the first version of the input syntax for MathML i discussed here,
> stuff as entitities was represented as empty elements. Not just for MathML
> but also for XHTML. E.g. &oacute; was <oacute/> and next defined via
> Schema (no DTD).

I don't see oacute as a math symbol; it's just another letter.
If it's used in one's locale, why not just use it as it is.
If it's not used in one's locale, it might be convenient to have
<ocaute/> in one's author-level documenttype, but I don't see it
as other than cdata for a browser-level documenttype.

> In a visual representation, you just copy tokens and the times is mainly
> redundant. But it is not for aural rendering. "4 times pi" is standard in
> spoken mathematics. [4 * pi] is better than [4 pi].

In my experience one usually says "4 pi" rather than "4 times pi".
Are you saying that mathml-capable aural user agents commonly do not
make a distinction between

      <mi>x</mi><mi>y</mi>    and    <mi>xy</mi>   ?

Perhaps a lame aural agent might fail in this regard.  But, on the
other side, a lame visual agent might also fail by inserting an
unknown symbol indicator for <mo>&invisibleTimes;</mo>.  (There is a
history of this.)

> Also the invisible times can be useful for automatic linebreaking.

But doesn't it bind rather tightly?  Inside an mrow wouldn't a + operator
usually be a better place to break?

As an author if I think the context for using <mi>x</mi><mi>y</mi>
might not be sufficient to imply invisibleTimes, then I will use
something explicit like \cdot.

                                    -- Bill

Received on Friday, 22 December 2006 17:02:16 UTC