Somehow I had the impression that this would be more than an overnight
The problem with those famous overnight jobs is that they tend to end
up with several weeks or month spent on something that is actually
secondary. I thought since there are already some free programs that do
render mathml, there should be some solution available.
But I guess we are going off topic here.
I just checked with the other people involved here and we would take up
Andrew Finney suggestion that we try to come up with a level A proposal
for Diagrams over the summer. That is if everybody agrees.
If this is the case, we would like to have some more feedback as to
what people think should be covered by such a level A proposal that is
not covered by the document we have put on the web.
In any case, we would constantly update our progress on this webpage in
order to see what everybody thinks and to see if we are moving in the
So here are some question that come to my mind:
a) If the above mentioned document would be updated to state that all
values are to be taken as being of unit pt (that is 1/72 (american (-;)
inch). Could we use this as a starting point?
b) Should we try to include render information like some basic shapes
into this level A proposal?
c) I agree the points raised by Eric Minch are important, but I also
think that this will not be a trivial task to nail that down. So should
this be addressed as well, or do we put it off till the other stuff has
been nailed down? Maybe Eric already has some idea how to include this?
On Montag, Mai 26, 2003, at 06:24 Uhr, Wagner,John wrote:
> Um...a really, really silly question, but...
> Why not just do it yourself?
> My experience with using other tools to accomplish
> what you want is, it's much easier to spend twice as
> much time finding and adapting the damn thing than to
> code it yourself, than it is to save half the time it would
> have taken you to code it yourself.
> So my general rule is, if you can find something off
> the shelf that solves exactly the problem you want to
> solve, then use it. Otherwise, do it yourself.
> Besides, if you've already got the expression in an
> expression tree, it's a night's work, maybe two.
> -----Original Message-----
> From: Ralph Gauges [mailto:firstname.lastname@example.org]
> Sent: Mon 5/26/2003 11:10 AM
> To: Wagner,John
> Cc: email@example.com
> Subject: Re: [sbml-discuss] DWG math display?
> Oh I fully agree. I would love to have some C library that renders
> MathML statement to a bitmap representation, but I haven't found one
> far. Actually I think I would already be happy with some tool that
> converts MathML to SVG, because I think there are libraries that
> SVG to bitmap, but besides some commercial product, I haven't found
> Actually this GtkMathView widget I emntioned can be used to just
> MathML to bitmap without actually having to use the GTK widget as your
> GUI toolkit, but you still need GTK installed because it is linked
> against it.
> JEuklid isn't an option because you need the whole tomcat and Cocoon
> order to use it.
> The mozilla people are also able to render MathML, but to find the
> that does it in all this code is certainly beyond me.
> I guess there are a lot of people who would like to be able to display
> some MathML as a nice looking graphic, so if anybody knows of some
> solution that is free, please step forward.
> As I said, right now, I convert all formulas to latex, run latex and
> convert the dvi file to png. It sure works and it looks a lot nicer
> especially for large formulas, but I am far from happy with it.
> The only thing I said was, that I do not agree with Herbert Sauro on
> the point that a java tool the renders MathML to some bitmap would be
> useless to programmers of other languages. I think that with newer gcc
> compiler you can make java classes into standard shared libraries for
> which you then can write wrapper for other languages.
> On Montag, Mai 26, 2003, at 04:11 Uhr, Wagner,John wrote:
> > I guess the way I was thinking about the whole math display is...
> > Either a program wants to display math in a nice and formatted
> > manner, or it doesn't. Those that don't want to display formatted
> > math simply won't display formatted math. But, those that *do*
> > want to display formatted math are faced with doing the drawing
> > and layout themselves. For *those* applications, it would be very
> > helpful to have the MathML, because it basically specifies the
> > layout for you.
> > For example, in the Virtual Cell, when you type in an expression
> > like (a+b)/(c+d/e), Jim Schaff has written a very nice algorithm
> > that takes the expression tree as input and outputs a "format tree"
> > very similar to the MathML display.
> > So I was thinking it might make it easier for those people wanting
> > to do something like that.
> > John
> > -----Original Message-----
> > From: Ralph Gauges [mailto:firstname.lastname@example.org]
> > Sent: Mon 5/26/2003 3:16 AM
> > To: Herbert Sauro
> > Cc: 'email@example.com '
> > Subject: Re: [sbml-discuss] DWG math display?
> > I fully agree that both options are far from optimal,
> especially the
> > second one since you need all those programs installed. On
> the other
> > hand, it seems to work as long as there is no other option.
> > Actually a java lib would not be that bad I think. Since the
> new gcc
> > compiler can compile java code to machine byte code, it
> should be
> > possible to create a normal library from java code and then
> write some
> > wrappers for perl, python, etc.
> > Maybe I am wrong. Has anybody else played around with the gcc
> > compiler yet?
> > Ralph
> > On Freitag, Mai 23, 2003, at 06:57 Uhr, Herbert Sauro wrote:
> > >
> > > All very well on paper but how do we programmers implement
> > Someone
> > > earlier mentioned that they only knew of two portable libs
> > that
> > > could render MathML, a GTK method and a long winded
> approach to
> > convery
> > > MathML to Latex, etc etc. Neither method sounds appleaing.
> > >
> > > I'm sure there is a Java lib somewhere but that isn't of
> much use to
> > > many of
> > > use who use other languages.
> > >
> > > Herbert
> > >
> > > -----Original Message-----
> > > From: Poul Nielsen
> > > To: Wagner,John
> > > Cc: Poul Nielsen; firstname.lastname@example.org
> > > Sent: 5/23/03 3:07 AM
> > > Subject: Re: [sbml-discuss] DWG math display?
> > >
> > >
> > > On Friday, May 23, 2003, at 14:56 Pacific/Auckland,
> > wrote:
> > >
> > >> There was also some discussion in Florida about display
> > >> MathML, and I got to thinking about it...it seems like we
> > >> could include display MathML markup in the Diagrams
> > >> portion. Imagine if, in addition to diagrams for the
> > >> reaction pathways, we also had a list of MathML display
> > >> elements that are linked to the various MathML expressions
> > >> in the SBML document.
> > >>
> > >> I was thinking of an addition to something like JDesigner
> > >> such that when you hovered the pointer over a reaction
> > >> arrow, you got a "tooltips" like action that popped up a
> > >> tooltip window, and in it was displayed that reaction's
> > >> kinetics; this MathML display element could provide the
> > >> layout for the content (just as an example).
> > >>
> > >> However, someone--maybe me?--poninted out that we
> > >> all pretty much knew how to do the math display anyway,
> > >> so it wasn't clear to me if there was any consensus that
> > >> display MathML was needed. Anyone care to weigh in
> > >> on this? Poul, any comments?
> > > MathML provides mechanisms to combine content and
> > markup.
> > > Chapter 5 of the MathML 2.0 specification is devoted to
> just this
> > > issue. It seems pretty straightforward to tie the two
> together to
> > > enable the behaviour that you describe above.
> > >
> > > Poul Nielsen
> > > Bioengineering Institute
> > > University of Auckland
> > > New Zealand
> > >
> > >