Re: mover vs latin chars with diacriticals

Andreas Strotmann and David Carlisle,

Thanks by your comments and corrections.

As said in

[http://lists.w3.org/Archives/Public/www-math/2006Apr/0075.html]

"I have not studied this in detail and cannot offer assistance."

It has been proven my lack of wisdom in this topic.

Once recognized that and after reading comments done by different authors
(and reading more about Unicode standard) I would add the following.

The question was either Unicode or <mover> and <munder>. Well,

It has been said that Unicode offers better rendering possibilities than
MathML presentational markup.

Some authors as Richard Kaye try by using Unicode when ü is an idiomatic
font and by ü when is a mathematical concept, for example second
derivative. Many people agreed in this general philosophy. Bruce Miller
did go beyond and suggested than something as

<mover><mi>q</mi><mo>.</mo></mover>

could be left to the operation on q, possibly a time derivative, whereas

<mover><mi>q</mi><mi>.</mi></mover>

would representing a single composed identifier.

I find this rather incompatible with general design of MathML. As Neil
Soiffer repeatedly stated, Presentational MathML is designed for notation
or presentation, whereas Content markup is for "meaning".

There is nothing in

<mover><mi>q</mi><mi>.</mi></mover>

or

<mover><mi>q</mi><mo>.</mo></mover>

or any other possibility denoting derivative. It is more, if we carefully
claim that Presentational MathML just render things as x^2 and x^2 without
disambiguating (this is the task of the <power/> tag of Content MathML for
example) then it may be difficult the justification for disambiguating ü
from ü in Presentation MathML 2.0.

To be more precise, ü, ü, and ü would be three rendered in the same way
(of course maybe with stylistic modifications).

Why would I use presentational markup for disambiguate ü (Spanish diaresis
u), from ü and ü (both derivative). And would I disambiguate the two
derivatives using some trick on Presentational markup or is this already
task for content markup, first being total second derivative whereas other
being local second derivative.

By using Presentational Markup we recover all difficulties due to
incompatibility of MathML with CSS, doing that css rules do not work in
MathML documents. This obligates to use two approaches: CSS for text and
other markup, whereas special <mstyle> tag for mathematical content.

Moreover, due to difficulties for unifying MathML layout with rest of
rendering engine on browsers, we can observe that even trivial style
commands (as those in the official MathML test suite) are not accepted
(try with Firefox for example) and one loses stylistic possibilities.
However, I can change style of text using css rules and this is in the way
that White Lynx illustrated here for q-dot.

Once says this, it may be remarked another points.

Using <mover> and <munder> we are using a specific markup model. This will
generate complications in the translation to/from other applications are
not based in MathML. By adopting Unicode, we are unifying the MathML and
not MathML tools on and off the web. Unicode has been designed for use on
so disparate systems as illustration programs, programming languages or
word processors. Using MathML specific code we are limiting usage.

Do not forget that Unicode is international standard whereas MathML is
just w3c recommendation. I think that rationale is to choosing more
general possible standard (over specific markup models) when available.
The rule would be (I believe) the more general and spreader over the less.

By using Unicode, we have a standard with little (or even none) variation
on final code. This not achieved via MathML support. White Lynx offered
different possibilities for rendering THE SAME in MathML. In fact, I
offered examples of MathML code generated by several MathML tools and
translated the code to another MathML tool (Mathematica 5.2 online tool)
for procesing.

That in TeX is encoded as \dot{q} in MathML was encoded in four different
ways. Two codes directly offered error in Mathematica 5.2 online
processing but when corrected by hand you can display them. From large
close points to small separated points passing by square block of width of
q.

One of promises of MathML was standardization of communication between
tools and authors. Something as simple as \dot{q} was presented in
different ways.

In fact, I believe on White Lynx exaggeration of one would find an
infinite number of presentational markup for THE SAME presentational
concept because writing \dot{q} (i.e. the same input code) in another tool
I obtained a different output from others I cited here in the past.

The translation of Presentational markup on Unicode for rendering may be
very difficult to achieve in practice because variability. Is it supposed
that browser would condense the "infinite" combinations of MathML
presentational code for q-dot in a single Unicode way?

That is, the author likes TeX and introduces \dot{q}, next each tool
generate a different MathML presentational markup, see

[http://lists.w3.org/Archives/Public/www-math/2006Apr/0084.html]

and

[http://lists.w3.org/Archives/Public/www-math/2006May/0003.html]

for posibilities. Next the browser engine would understand that all those
combinations (more others programmer may imagine could be offered) are the
same (from a presentational point of view) and just render them as Unicode
combined diacriticals.

Is that?

Moreover Unicode is also designed for search and this would help to search
engines to match.

Note the requirements and points illustrated here. Whereas I almost agree
with last Stan Devitt post

[http://lists.w3.org/Archives/Public/www-math/2006May/0010.html]

I think that he has missed a bit the point when says

<blockquote>
I am hearing an argument  to throw out presentation MathML because it
fails an implicit requirement that "a particular mathematical concept be
uniquely represented" and to use a simple character based representation
because it is closer to a unique representation and so easier to search
for.

1.  This argument ignores the fact that notations get re-used for
different concepts in mathematics.  To search reliably for a concept, at
very least you need the kind of information contained in the expression
</blockquote>

In above examples, one is not comparing different notations; one is
comparing THE SAME notation but expressed in different ways in
presentational MathML markup.

<blockquote>
2.  It is unreasonable to  expect that a single concept to be "presented"
uniformly by all authors or applications (even as a multi-character
string) unless perhaps the presentation is generated by the same
author/system on
the same day or is machine generated by the same software.  Even then,
other cultures may deliberately choose a different presentation.
</blockquote>

Maybe a full complete unification (an only way) was impossible, but note
all outputs were generated from the same input: \dot{q}.

Two or three representation are preferable to a dozen. Note also that
Unicode define ways to compare different codes.


Juan R.

Center for CANONICAL |SCIENCE)

Received on Wednesday, 3 May 2006 09:50:37 UTC