RE: pages with MathML

Paul Libbrecht wrote:
>
> Juan,
>
> I see you are aiming to expect from classical authors that produce TeX
> sources that they provide a fully semantically-correct formula.

TeX is limited, I am not waiting that.

I just do not want to see people claiming that they have "the world’s more
advanced site" or that they are adopting MathML because is accessible or
similar ones and after something as simple as (ds)^2 (is not?) is being
encoded and served in MathML as d(s)^2 (HERMES is not the one doing
that!).

Using MathML for simulating prescripts via

<msup><mrow/><mi>B</mi></msup><mi>F</mi>

is more ridiculous than using HTML for that. Using GIFS for

\int sin(x) dx

with the "standard" Nemeth code for aural rendering "integral of integrand
sin of x dx" on ALT attribute) is more accessible that via "modern" MathML
code is being generated and distributed by blogs, websites, and journals I
cited and others I am not citing.

> I wish you would be right but I can assure you that making it possible
> for authors to be both satisfied with the presentation of their content
> as well as with the underlying semantic is a challenge that is current
> research. And what could even be more of a challenge is to let the
> authors realize what the semantic could mean to them!

Please note that I focused on presentation MathML not in content.

If you want simply visual rendering of a^b you can do that via old HTML

<span>a</span><sup>b</sup>

or using some other approach. None of us wait to see spreading in the
Internet incorrect encodings like

<mi>a</mi><msup><mrow/><mi>b</mi></msup>

If really we are receving that incorrect piece of code, then it is better
to return to the old HTML and do not worry people to update to modern
browsers, download and install special fonts or plugins...

In above -and others cases- the old HTML (which was not designed for
mathematics) is encoding better the structure of math, render better, and
is being more accessible than using specific MathML. People would think a
bit about that before was too later.

Part of bad MathML code is being served in the Internet is directly
related to weakness of the MathML specification and that I call "design
errors".

I was a MathML believer once, how wrong I was! Now, I consider that
approaches as XML-MAIDEN are better for a future mathematical/scientific
web. That is reason I abandoned MathML.

XML-MAIDEN is more accessible, better structured, compatible with DOM and
CSS, and can be rendered in almost any browser without special fonts,
plugins... Moreover, George has designed scientific semantic layers (do
not existing in MathML) and the code can be edited by hand (letting you
fine-tunning).

I am perplexed that using a simple text editor and a *standard* CSS
compatible browser, George is able to render scientific content I am
unable to see using Firefox (with native MathML), restricted special DTD
(XHTML+MathML), and specialized tools (needed for type the verbose MathML
code).

Of course, I understand the effort the many people did in MathML. But
please, note that I already did that effort for providing MathML support
in canonicalscience website and in the future Canonical Science Report
journal. Therefore I can understand you guys!

I simply have found insurmountable difficulties and abandoned the
CanonMath project. I have discussed a couple of questions with George and
he has corrected several errors mine. Now I think that final code will be
generated and server at the Center will be something very similar to
George’s XML-MAIDEN.

George’s markup is very close to real mathematical/scientific content and,
therefore, it will be easier for authors to find errors in both semantic
and presentation layers.

A mathematician writing d(s)^2 knows that he|she is typing wrong and
correct it to (ds)^2. If you are writing the code via a tool or in
semi-tricky TeX and you are trained to not see the underlying MathML code
(or if you want see it but you cannot understand it if you never studied
MathML language) then you cannot know if you are really typing d(s)^2 or
d(s)^2 or any else.

There are lot of errors in real MathML is spreading on the Internet. One
very amazing is from Distler’s blog that during last years has served an
unknown quantity of code such as <mn>3<mn><mo>.<mo><mn>1416<mn> for
3.1416.

Is this the future of mathematical communication on the web?


> paul
>

[snip]


Juan R.

Center for CANONICAL |SCIENCE)

Received on Saturday, 15 April 2006 14:50:32 UTC