Subject: Re: gtk MathView.
From: Martin Sevior (email@example.com)
Date: Mon Sep 18 2000 - 06:47:47 CDT
On Mon, 18 Sep 2000, Luca Padovani wrote:
> Hi Martin,
> > 1. is a lot more work for our already strained resources but it retains
> > our cross platform capability. 2. requires that our gnome front end
> > becomes a Gnome container but we can leverage other peoples code much more
> > effectively at the expense of cross platform capability. ie Only the Gnome
> > front end will be able to display equations.
> at present this is true, but the widget is designed to be "device
> independent". In theory one has just to write 3 classes to get it
> working with another graphic interface or whatsoever. In fact, I'm just
> planning to do this in order to output PostScript. The 3 classes are
> DrawingArea (with drawing primitives as DrawLine, DrawString...),
> GraphicContext (the name says all) and (obviously) Font. The last one
> might be the trickiest, due to the lack of a uniform font management in
This is really cool. It sounds like it would integrate into the Abi cross
platform abstraction very easily. These functions are all provided by
xp virtual functions that are subclassed into platform specific functions.
It might be easy to incorporate your code into Abi at a cross platform
level. In which case you automatically get Windows, BeOS, QNX and Gnome
front ends for your widget. We get MathML on all our platforms. So
it's Win-Win :-)
> > 1. It notifies GtkMathML. > > 2. Feeds MathML to GtkMathML (which
> > 3. Asks GtkMathML for the size of the equation in height and width.
> > 4. If needed it asks to break the equation at a certain width.
> well, at the moment linebreaking doesn't work this way. The problem is
> that the widget is responsible for finding a suitable point to break a
> long expression, this point is computed looking at the structure of the
> expression but may be influenced by specific MathML elements. Moreover,
> it is the "environment" (in our case would be AbiWord) which feeds the
> widget with the space available to render the expression (so it is a
> top-down process rather than a bottom-up process). IMHO, it won't be so
> easy to integrate the widget taking into account linebreaking, at least
> without some extended modification. However, this has to be eventually
> done to make integration effective.
This is not a problem at all. We can tell the widget when to break.
> > It will be much harder to come up with a GUI to input MathML. I have some
> > ideas but they are rather nebulous. I have only made a cursory
> > investigation of DOM technology and it sounds great. However there seems
> > to be a distinct lack of progress in real world usage of DOM.
> Well, I'm not a WYSIWYG fanatic and all this makes me a bit worried about
> performances, because these "beautiful" technologies, once implemented,
> are adding more and more levels of abstrasction between applications and
> the stored data. Maybe this will serve as justification for 1Ghz
> processors... :-)
Speed is not an issue for the GUI I think. Unfortuntely it will be a big
challenge to enable the user to enter maths quickly :-)
> > Do you mind if I post this email to the Abiword and gnome-office mailing
> > lists? I think they would be interested in this discussion.
> it's ok.
This archive was generated by hypermail 2b25 : Mon Sep 18 2000 - 06:48:12 CDT