"I spoke personally with Chris Hoffman of Mozilla Foundation about
MathML and he told me that the project leader bailed out for some
other project a while ago. This means the project is basically
stalled. The project page was last modified September, 2006, and most
links on that page are broken."

There's been a good bit of work on MathML since then to keep it
working after the switch to cairo and the reflow branch landing
(mostly by Karl Tomlinson, I think).  I think this means your
message's subject is an overstatement.

However, I don't think there's been much, if any, new feature work
on MathML, although I might be missing something.

-David

--
L. David Baron                                 http://dbaron.org/
Mozilla Corporation                       http://www.mozilla.com/

On Thu, Aug 20, 2009 at 7:33 AM, L. David Baron<dba...@dbaron.org> wrote:
> On Wednesday 2009-08-19 19:39 -0700, net-buoy wrote:
>> Thought this bit of news as posted on moodle.org where some folk are
>> trying to promote use of mathml via asciimathml and other options was
>> worth sharing:
>>
>> "I spoke personally with Chris Hoffman of Mozilla Foundation about
>> MathML and he told me that the project leader bailed out for some
>> other project a while ago. This means the project is basically
>> stalled. The project page was last modified September, 2006, and most
>> links on that page are broken."
>
> There's been a good bit of work on MathML since then to keep it
> working after the switch to cairo and the reflow branch landing
> (mostly by Karl Tomlinson, I think).  I think this means your
> message's subject is an overstatement.
>
> However, I don't think there's been much, if any, new feature work
> on MathML, although I might be missing something.

That's unfortunate.  There is plenty work that could and should be done
with the MathML support in Firefox--including accessibility work.

I thought Mozilla was looking to revive the project and apply a bit
of resources to MathML enhancements.

--
--Alex Milowski
"The excellence of grammar as a guide is proportional to the paucity of the
inflexions, i.e. to the degree of analysis effected by the language
considered."

Bertrand Russell in a footnote of Principles of Mathematics

> On Thu, Aug 20, 2009 at 7:33 AM, L. David Baron<dba...@dbaron.org> wrote:
>> On Wednesday 2009-08-19 19:39 -0700, net-buoy wrote:
>>> Thought this bit of news as posted on moodle.org where some folk are
>>> trying to promote use of mathml via asciimathml and other options was
>>> worth sharing:
>>>
>>> "I spoke personally with Chris Hoffman of Mozilla Foundation about
>>> MathML and he told me that the project leader bailed out for some
>>> other project a while ago. This means the project is basically
>>> stalled. The project page was last modified September, 2006, and most
>>> links on that page are broken."

Walter's comment kind of needs to be read with the knowledge that
he is advocating an alternative technology.

But, as Walter indicates in another comment, even though the
volunteers that have done a lot of work on MathML in Mozilla in
the past have not been active recently, that does not mean that
Mozilla is not interested.

>> There's been a good bit of work on MathML since then to keep it
>> working after the switch to cairo and the reflow branch landing
>> (mostly by Karl Tomlinson, I think).  I think this means your
>> message's subject is an overstatement.
>>
>> However, I don't think there's been much, if any, new feature work
>> on MathML, although I might be missing something.
>
> That's unfortunate.  There is plenty work that could and should be done
> with the MathML support in Firefox--including accessibility work.

Frédéric Wang recently added the menclose feature:
https://bugzilla.mozilla.org/show_bug.cgi?id=297467

and Mozilla would welcome other contributions.

> I thought Mozilla was looking to revive the project and apply a bit
> of resources to MathML enhancements.

It takes resources just to keep MathML in Gecko while changes are
made to Gecko's infrastructure and to ensure that supporting
MathML doesn't compromise the safety of using a browser.
Mozilla has been applying resources here.

Yeah, we should update the pages to reflect reality. Maybe Karl can do that.

Mozilla staff are currently maintaining our MathML support (i.e.,
keeping it working as other things change around it) and willing to
review and land contributions to improve it, but we aren't going to
invest in any major improvements until MathML becomes a lot more popular
on the Web. We already support a very large subset of MathML and there's
just no point in extending that until people start using what we've
already implemented.

Actually there is one major improvement for MathML coming as a side
effect when we transition to the new HTML5-spec-based parser; you will
be able to write MathML in non-XML HTML documents. That will make MathML
far more usable on the Web.

Rob

On Sun, Aug 23, 2009 at 4:02 PM, Robert O'Callahan<rob...@ocallahan.org> wrote:
>
> Mozilla staff are currently maintaining our MathML support (i.e., keeping it
> working as other things change around it) and willing to review and land
> contributions to improve it, but we aren't going to invest in any major
> improvements until MathML becomes a lot more popular on the Web. We already
> support a very large subset of MathML and there's just no point in extending
> that until people start using what we've already implemented.

That seems somewhat circular in that until there is a high quality
implementation
people won't use it.  Using MathML is currently quite the "magic incantation"
that also requires a user to install fonts.  Most users won't do that.

> Actually there is one major improvement for MathML coming as a side effect
> when we transition to the new HTML5-spec-based parser; you will be able to
> write MathML in non-XML HTML documents. That will make MathML far more
> usable on the Web.

Yes, that will certainly be a big help with the deployment of MathML by
authors.

--
--Alex Milowski
"The excellence of grammar as a guide is proportional to the paucity of the
inflexions, i.e. to the degree of analysis effected by the language
considered."

Bertrand Russell in a footnote of Principles of Mathematics

On 23/08/09 7:41 PM, Alex Milowski wrote:
> On Sun, Aug 23, 2009 at 4:02 PM, Robert O'Callahan<rob...@ocallahan.org>  wrote:
>> Mozilla staff are currently maintaining our MathML support (i.e., keeping it
>> working as other things change around it) and willing to review and land
>> contributions to improve it, but we aren't going to invest in any major
>> improvements until MathML becomes a lot more popular on the Web. We already
>> support a very large subset of MathML and there's just no point in extending
>> that until people start using what we've already implemented.
>
> That seems somewhat circular in that until there is a high quality
> implementation
> people won't use it.  Using MathML is currently quite the "magic incantation"
> that also requires a user to install fonts.  Most users won't do that.

I thought that that problem is going away since Windows Vista and the
latest version of Microsoft Office include Cambria Math, Linux distros
will package STIX when that's released, and Mac already had adequate
fonts. Is that not true?

I also believe that in Firefox 3.5 you could put a free math font on
your server and use CSS @font-face to instruct the browser to download
and use the font.

Rob

Alex Milowski <al...@milowski.com> writes:
> On Sun, Aug 23, 2009 at 4:02 PM, Robert O'Callahan<rob...@ocallahan.org> wrote:
>>
>> Mozilla staff are currently maintaining our MathML support (i.e., keeping it
>> working as other things change around it) and willing to review and land
>> contributions to improve it, but we aren't going to invest in any major
>> improvements until MathML becomes a lot more popular on the Web. We already
>> support a very large subset of MathML and there's just no point in extending
>> that until people start using what we've already implemented.
>
> That seems somewhat circular in that until there is a high quality
> implementation
> people won't use it.

It's not the quality of the Mozilla implementation, that stops publishers from
using MathML. But you have to offer an other representation of math, for example
images,for those people, who use other browser. Then, if the other
representation works well im Mozilla too, why would you maintain two
representations?

Martin

On Mon, Aug 24, 2009 at 1:41 AM, Martin Holz<usene...@martin-holz.net> wrote:
> Alex Milowski <al...@milowski.com> writes:
>
> It's not the quality of the Mozilla implementation, that stops publishers from
> using MathML. But you have to offer an other representation of math, for example
> images,for those people, who use other browser. Then, if the other
> representation works well im Mozilla too, why would you maintain two
> representations?

It is great to have some kind of implementation of MathML and I am certain
quite happy that Firefox contains some support for MathML.  Unfortunately, that
support is incomplete and still requires addition fonts on most deployed
platforms.  The coming STIX fonts will help alleviate the font problem but
it is unclear how Mozilla will achieve uniformity given OS vendors release
cycle and commitment to including the STIX fonts.  Most certainly, people
with older OS version (of which there are many) will not have them.

Having to use CSS and have the browser download a font isn't very friendly
to a publisher and most will attribute that to quality of
implementation--however
fair or unfair that is.

That's where I come back to active support by Mozilla of the "big three" of
MathML, SVG, and HTML.  Just maintaining the code isn't sufficient.

--
--Alex Milowski
"The excellence of grammar as a guide is proportional to the paucity of the
inflexions, i.e. to the degree of analysis effected by the language
considered."

Bertrand Russell in a footnote of Principles of Mathematics

On Aug 24, 1:02 am, Robert O'Callahan <rob...@ocallahan.org> wrote:
>
> Mozilla staff are currently maintaining our MathML support (i.e.,
> keeping it working as other things change around it) and willing to
> review and land contributions to improve it, but we aren't going to
> invest in any major improvements until MathML becomes a lot more popular
> on the Web. We already support a very large subset of MathML and there's
> just no point in extending that until people start using what we've
> already implemented.

MathML is becoming more and more popular. For examples see
http://dlmf.nist.gov/
http://www.firemath.info/

It would be nice if Mozilla would not only follow the mainstream (add
video support etc.) but feel responsible of supporting science
(minorities) as well. In the end many of the technologies used within
Firefox came out of science. So it would be only fair to give some
dividend back.

Firemath wrote:
> MathML is becoming more and more popular. For examples see
> http://dlmf.nist.gov/
> http://www.firemath.info/

        Nice collection of formulae. Wikipedia is certainly one of the main
resource for mathematics on the Web but unfortunately is still using PNG
images... http://www.mediawiki.org/wiki/Extension:Blahtex

>
> It would be nice if Mozilla would not only follow the mainstream (add
> video support etc.) but feel responsible of supporting science
> (minorities) as well. In the end many of the technologies used within
> Firefox came out of science. So it would be only fair to give some
> dividend back.

        I've just gave a try to your extension. It's certainly useful to
quickly insert mathematical symbols and write simple presentation
MathML. However It's not easy to generate a complex recursive structure
(it often says "Operation not permitted on empty objects!"). I also feel
difficult to move across the tree structure and select fragments of the
formula. It lacks a parsing for simple expression such that "a+b=2" and
copy&paste feature of MathML fragments.
        I hope one day I can hack mozilla/source/editor/ to add the editing
features of Amaya (including the recent SVG editor BTW) so that XUL
applications like yours (maybe Thunderbird?) can rely on a user-friendly
interface...

On 11/09/09 1:05 AM, Firemath wrote:
> It would be nice if Mozilla would not only follow the mainstream (add
> video support etc.) but feel responsible of supporting science
> (minorities) as well. In the end many of the technologies used within
> Firefox came out of science. So it would be only fair to give some
> dividend back.

Well, we have been supporting MathML for years longer than everyone else
--- and that hasn't been free, belive me.

Rob

I know that, personally, I really appreciate that Mozilla has been supporting
MathML as it exists in the code base.  I think what you are seeing is that
many of us want more out of the implementation and that just maintaining
the current state without some forward movement is something we
all have opinions about.

With the current development of HTML5 going on, it would be really
great to see HTML5 progress in conjunction with SVG and MathML as
a fully supported triad of markup we can rely upon in Firefox and other
Mozilla derived products.

Maybe we should turn this discussion in a positive direction:

   1. What feature(s) of MathML 2 do you need that is not implemented?

   2. What's "broken" that you need fixed?

--
--Alex Milowski
"The excellence of grammar as a guide is proportional to the paucity of the
inflexions, i.e. to the degree of analysis effected by the language
considered."

Bertrand Russell in a footnote of Principles of Mathematics

> Firemath wrote:
>> MathML is becoming more and more popular. For examples see
>> http://dlmf.nist.gov/
>> http://www.firemath.info/
>
>     Nice collection of formulae. Wikipedia is certainly one of the main
> resource for mathematics on the Web but unfortunately is still using PNG
> images... http://www.mediawiki.org/wiki/Extension:Blahtex

Yes, Wikipedia has a potential for boosting the use of MathML with its
tons of equations. It is supposed to have experimental functionality to
use MathML for formulas: you have to login with a personal account, and
in your settings for Math select the option "MathML where possible
(experimental)". However, I don't see any effect: A few simple formulas
get converted to plain HTML, and most of them to PNG.

I think, there are libraries for converting LaTeX (the internal math
representation in Wikipedia) to MathML. They should implement it and get
rid of the rasterised PNGs, at least for users of Firefox :-)

[OT] I remember I also suggested Wikipedia to embed SVG drawings as such
(they currently convert them to PNG for display), for SVG-enabled
browsers, but they were reluctant as they considered SVG to be
insufficiently or differently implemented in the various browsers.

On 16/09/09 10:27 AM, Alvaro wrote:
> Yes, Wikipedia has a potential for boosting the use of MathML with its
> tons of equations. It is supposed to have experimental functionality to
> use MathML for formulas: you have to login with a personal account, and
> in your settings for Math select the option "MathML where possible
> (experimental)". However, I don't see any effect: A few simple formulas
> get converted to plain HTML, and most of them to PNG.
>
> I think, there are libraries for converting LaTeX (the internal math
> representation in Wikipedia) to MathML. They should implement it and get
> rid of the rasterised PNGs, at least for users of Firefox :-)
>
> [OT] I remember I also suggested Wikipedia to embed SVG drawings as such
> (they currently convert them to PNG for display), for SVG-enabled
> browsers, but they were reluctant as they considered SVG to be
> insufficiently or differently implemented in the various browsers.

It would be great if someone found out what their specific issues are,
and why their MathML option isn't working!

Rob

> It would be great if someone found out what their specific issues are,
> and why their MathML option isn't working!
>
> Rob

There exists an extension for MediaWiki called BlahTex, and it seems
ready for a while.

        http://www.mediawiki.org/wiki/Extension:Blahtex
        http://gva.noekeon.org/blahtexml
        
The original developer is David Harvey. Apparently, he stopped leading
the project because the core developers of MediaWiki have never taken a
moment to integrate it. It is now maintained by Gilles van Assche. See
this discussion:

http://en.wikipedia.org/wiki/Wikipedia_talk:WikiProject_Mathematics/Archive_14#MathML_.2F_improved_TeX_support

 From this discussion, a new bug was created to request integration of
Blahtex in future versions of MediaWiki:

https://bugzilla.wikimedia.org/show_bug.cgi?id=6383

This bug is three years old but the developers of MediaWiki still have
not answered. I have just voted for this bug, maybe gathering enough
votes will catch their attention.

While google chart API prsents some new options, it is high time that
google allow mathml.  I do not have google script access and wonder if
script access can be used to invoke asciimathml.js

btw, my original posting was done rather tongue in cheek.....   But it
would be nice if mathml in general took a giant step forward ;)

On Sep 19, 9:30 am, Frédéric Wang <fred.w...@free.fr> wrote:
> > It would be great if someone found out what their specific issues are,
> > and why their MathML option isn't working!
>
> > Rob
>
> There exists an extension for MediaWiki called BlahTex, and it seems
> ready for a while.
>
>        http://www.mediawiki.org/wiki/Extension:Blahtex
>        http://gva.noekeon.org/blahtexml
>
> The original developer is David Harvey. Apparently, he stopped leading
> the project because the core developers of MediaWiki have never taken a
> moment to integrate it. It is now maintained by Gilles van Assche. See
> this discussion:
>

>
>  From this discussion, a new bug was created to request integration of
> Blahtex in future versions of MediaWiki:
>
> https://bugzilla.wikimedia.org/show_bug.cgi?id=6383
>
> This bug is three years old but the developers of MediaWiki still have
> not answered. I have just voted for this bug, maybe gathering enough
> votes will catch their attention.

        About mixing formulas & graphics, the W3C has not specified entirely
how this should be done, even if MathML & SVG specs and note [3] give
some hints. The inclusion of a formula inside a graphics should be
reflected by the DOM structure: the MathML tree should be a subtree of
the SVG one. Hence the use of a <foreignObject/> to put MathML inside
SVG (again, see [2]) is better than using overlapped CSS <div/>'s. For
instance, it is expected than a copy&paste of the <svg/> also acts on
the formula and this is straightforward when <foreignObject/> is used.
Another example, for people who have Amaya or a SMIL-enabled build of
Mozilla: a formula *really* included inside SVG can be transformed
and/or animated [4].

        Back to the chicken or the egg dilemma, I think that implementers
should take the first step and propose as much features of the Spec as
possible. People do want to use formulas on the Web and if the current
implementations are not sufficient, they are unfortunately going to use
workarounds such that png/svg images or javascript-generated MathML DOM
+ CSS. Thanks to the switch to Cairo and Karl's work, MathML-in-SVG is
now possible in Mozilla. Allowing MathML/SVG in HTML is another step
forward, since it seems that the current requirement that a document
must be served as XML is one of the main limitation for users. Finally,
another improvement in Mozilla that has not been previously given:
Thunderbird is now built with MathML support [5].
        
[1] http://search.mathweb.org/index.xhtml
[2] http://golem.ph.utexas.edu/instiki/show/Sandbox
[3] http://www.w3.org/Math/Documents/Notes/graphics.xml
[4]
http://www.maths-informatique-jeux.com/international/mathml_with_other_standards/mathml_svg.xml
[5] https://bugzilla.mozilla.org/show_bug.cgi?id=456014

I appreciate comment about formula inside graphic, as with asciimathml
I am still surprised by some of the responses (one artifact is in fact
explored in the edtech pages I cited ;=}  There are few "perfect"
solutions, but if I can get a student to use asciimathml syntax to
express equations so that they are rendered in mathml without needing
a server or expensive client side software like Scientific Notebook,
isn't that worthwhile ;=}

The Moodle Guy (I am going to assume you are talking about Mauno as
opposed to me, though we are both active on Moodle.org)  and I may
well agree with you on quite a few points that you make, but if we
want mathematics notation to be transparent and universal, we can't
always expect that the user has a server side tex installation to rely
on..... and what guarantee that every web 2.0 app will employ the same
server side technology?  So,  while I don't want to see Firefox become
a huge juggernaut, I would love to see firefox be able to
transparently parse tags from any of a half a dozen different text
syntax into MathML for display.....  which argues client, not server
side,  parsing and display....

In other words, there are if you will, two stages involved in display
of Math as few will ever write math in MathML; text expression to
MathML, then MathML to display. We need to address both stages.  I
think we also need to be inclusive of those using non-Tex syntax.  We
woud perhaps need new firefox extensions for libs needed and a tag for
stage one math to differentiate it from current mathml tag,  or
perhaps extensions to Mathml to address this??  Now,  someone could
for example do <mathtext>$\root{n}{x}$<\mathtext> and feel confident
that a) the token used "$" will be recognized as a Tex token and b)
based upon the use of a Tex token in that span the span will be
displayed as MathML using a tex2Mathml routine.

Vis-a-vis asciimathml on wikipedia,  as I demonstrated on edtech,
asciimathml can be implemented on mediawiki via theme (i.e. without
extension) and this can be replicated on other wikis (even dokuwiki
with some mods that have been documented) which offers some intriguing
options, and as Mauno has been demonstrating over the past week, the
Google Chart API can be used for fallback.

I think we need to not only have our eye on the pie in the sky,  but
we need to agree on some baby steps in that direction....  and we need
to make all this accessible for the teacher facing 150 tweens every
day.

Marc

p.s. And thanks for the work you have been doing    ;=}

> [4]http://www.maths-informatique-jeux.com/international/mathml_with_othe...
> [5]https://bugzilla.mozilla.org/show_bug.cgi?id=456014

> if we want mathematics notation to be transparent and universal, we
> can't always expect that the user has a server side tex installation
> to rely on..... and what guarantee that every web 2.0 app will employ
> the same server side technology?  So, while I don't want to see
> Firefox become a huge juggernaut, I would love to see firefox be able
> to transparently parse tags from any of a half a dozen different text
> syntax into MathML for display.....  which argues client, not server
> side, parsing and display....

I humbly but strongly disagree.  This is an example of how the 'net has
been going the wrong way: Instead of a few, clean specs we have a
proliferation of competing technologies, many proprietary, that are
supported to various degrees by different platforms.  Instead of the
one-markup-fits-all dream behind the separation of content from
rendering, most Web pages are highly platform specific. Expecting
Mozilla to support different math content formats would go down the same
road, making things worse.

MathML is *the* format for marking up math. If it falls short, let's
improve it. May all browsers do their best to implement it.

How MathML is generated is none of Mozilla's business.  There's a dozen
ways to do it. The choice is on the content creator's side, and this is
a good thing.

Let's keep these two separate: The server does its best to produce clean
markup, and the client does its best to produce a clean rendering from
this markup. The cleaner the set of markups to support by either side,
the better for everyone.

Justus

Devils advocate wise, mathml is a w3c standard and we can see the
impact that has had on Microsoft and webkit....  Wolfram is not going
to abandon mathematicas syntax and the same applies to
thosecomfortable with maxima or any other syntax.....  The problem is
that the paradigm of web content provider vs consumer breaks down with
"web 2.0" and will arguably disappear as web service takes over
webinteraction (web 3.0). If I amcomfortable using asciimath
ingooglegroups and you arecomfortable with maxima, why shouldn't we
bothbe able to use the syntax we arecomfortable with, having our
browsers adjust?

I think our real choice is between browser based or reliance on 3rd
party servers (such as WA, googlechartapi, or even mathtran or
mimetex) and sitting offline reading a file via my browser why
shouldn't I expect what is imminently possible?

On Oct 6, 8:18 am, Justus-b...@Piater.name wrote:
> net-buoy <marc.gro...@gmail.com> wrote on Mon, 5 Oct 2009 14:43:04 -0700

> Apologies, Justus, but I think you missed my point.
> I am not arguing for any dilution of mathml, just the opposite. I am
> arguing for additional tag standards, either as a superset to mathml
> or to XHTML that would allow the tagging of text expression for
> parsers.
> Likewise I am not arguing that FF do it all, merely that the extension
> architecture combined withsuch a tag provides the transparency and
> universality that is needed; ie parsers for text syntax can be added
> and turned on and off.

I agree with Justus here. Alternative input syntax for mathematics are
best thought of like wiki syntax for natural language text. Wikis became
popular because they didn't require people to install special browsers,
the conversion from the linear syntax to html/xml markup happens on the
client. Adding multiple math language parsers to a browser would be like
installing parsers for every flavour of wiki syntax.

> Devils advocate wise, mathml is a w3c standard and we can see the
> impact that has had on Microsoft

Quite a lot on microsoft, although not currently in IE, Word has MathML
input/output as standard and the new pen based math input in Windows 7
is using mathml internally as well as on file export.

>  and webkit....

it's comming....

https://bugs.webkit.org/show_bug.cgi?id=29529

>  Wolfram is not going to abandon mathematicas syntax

why should they? Mathematica has had MathML import/export since the
first drafts of MathML.

> I think our real choice is between browser based or reliance on 3rd
> party servers (such as WA, googlechartapi, or even mathtran or
> mimetex) and sitting offline reading a file via my browser why
> shouldn't I expect what is imminently possible?

A plugin to the browser (or javascript or whatever) that expands markup,
whether it be asciimath to mathml, or wiki markup to html, is useful
agreed, but the strength of the web is in sharing common markup
languages, not that everyone has to have a parser for every language
variant.

David

________________________________________________________________________
The Numerical Algorithms Group Ltd is a company registered in England
and Wales with company number 1249803. The registered office is:
Wilkinson House, Jordan Hill Road, Oxford OX2 8DR, United Kingdom.

This e-mail has been scanned for all viruses by Star. The service is
powered by MessageLabs.
________________________________________________________________________