Re: multi-symbol variables

Hi all,

1. Indeed, if we were encoding Content MathML, the B' example could
have been an easy <msup intent="ci">...</msup>

2. For AT, how does one know the words to speak B' ? And the better
example David Farmer gave was B'' spoken with "B-double-prime" in
 - I align with Murray's answer that if the narration overlaps the
baseline presentation tree+Unicode readout, no annotation is needed.
 - and when the narration requires a special treatment, writing it in
the intent attribute seems workable. We essentially overlap with the
aria-label technique, as in <msup intent="B-double-prime">

3. The embellishments are open-ended and tend to get more involved as
we get into higher math. Often a single name will have a subscript
*and* a superscript, or some combination of under,over,sub and super.
A random article from arXiv to illustrate would be:

Say taking, Definition 2.1 (4) here, we face the TeX
and the implied MathML:

<munder accentunder="true">

Hopefully this is spoken "p-n-tilde", but I am unsure. This document
alone has about 260 uses of \underset for using this embellishment

There are also more semi-related examples in that one PDF:
- a sub-cascade with alef:  א-sub-א-sub-א-sub-0 , which blurs the
distinction between expression and name, since yes - it stands for a
very concrete sequence of operations. But, when you communicate the
term, you don't actually go into explaining how the operations work.
One may even shorten to "the third" or "the triple" aleph zero when
narrating, if the context allowed it.

- embellishments with primes and stars are also present in this same document.
- Some compound names too, as in F^{\tau *}, maybe pronounced
"F-tau-star". Implied MathML:


Plenty of diversity.


On Fri, Oct 8, 2021 at 3:43 AM David Carlisle <> wrote:
> On 08/10/2021 04:09, David Farmer wrote:
> >
> > Taking this discussion out of GitHub.  I hope this is the right
> > place.
> >
> > The important point which has arisen, which I believe we had not
> > discussed previously, is how to mark a multi-letter string as
> > representing a single variable.  For example, on this page:
> >
> >
> >
> >
> > it says
> >
> > Interest = p \times r \times t .
> >
> > Here "Interest" is an 8 character string which represents one
> > quantity.
> >
> > This is analogous to the situation in our recent discussions,
> > where B' is just one quantity which happens to be a string of
> > two characters.  The fact that one of the characters is a prime
> > is not the main issue.  I mentioned an example where a_1 was a
> > (subscripted) string which also is really just a single variable.
> >
> > Prior to this week I don't think we had discussed the fact that
> > B' is a 2-character quantity, while f' -- assuming that means the
> > derivative of f -- is the object f on which ' is acting.
> >
> > The only point I am trying to make here is that we need a way to
> > indicate that a collection of characters is really in indivisible
> > unit which represents a single quantity.  The recent discussions were
> > made more difficult by the fact that this had not been discussed
> > previously.
> >
> > I suggest that " B' " should be handled the same way as " Interest "
> > would be handled in the formula above.
> >
> > Regards,
> >
> > David Farmer
> >
> logically I agree that "B'" is the same as "Interest" but typographic
> details don't always follow logic,.
> Interest is <mi>Interest</mi> which does the right thing automatically
> due to the rule about multi-letter mi using upright font. So it is easy
> in that case to say the default intent is the single name "Interest"
> B'  is <msup><mi>B</mi><mo>&prime;</mo></msup> or arguably
> <msup><mi>B</mi><mo>'</mo></msup> or <mi>B'</mi>
> Only the last of these has a natural intent as a single identifier (and
> traditionally mathml has been explicit that that form is not correct)
> mathml's traditional solution to this was to allow presentation in
> content mathml token elements so
> <ci><msup><mi>B</mi><mo>&prime;</mo></msup></ci>
> is unambiguously a variable name B' but we are trying to do without
> content mathml here so..  I think the natural intent to use here would be
> <msup intent="B prime"><mi>B</mi><mo>'</mo></msup> if the default intent
> on <mi>Interest</mi> is intent="Interest"
> or
> <msup intent="variable-name"><mi>B</mi><mo>'</mo></msup> if the default
> intent on <mi>Interest</mi> is intent="variable-name" (or some other
> known keyword to be decided) that means it's a variable and has natural
> reading just by reading the content.
> David
> Disclaimer
> The Numerical Algorithms Group Ltd is a company registered in England and Wales with company number 1249803. The registered office is: 30 St. Giles, Oxford, OX1 3LE, United Kingdom. Please see our Privacy Notice for information on how we process personal data and for details of how to stop or limit communications from us.
> This e-mail has been scanned for all viruses and malware, and may have been automatically archived by Mimecast Ltd, an innovator in Software as a Service (SaaS) for business.

Received on Friday, 8 October 2021 12:06:48 UTC