Subject: Re: FAQ -- equation editor
From: Paul Rohr (paul@abisource.com)
Date: Tue Feb 29 2000 - 19:11:19 CST
At 10:45 AM 3/1/00 +1100, Martin Sevior wrote:
>Could you put me in touch with the developer? This is the next thing I
>want to work on. I know Justin is also keen to work on this too.
Done.
>To this point, I know very little about XML and how the fields issues will
>be resolved. However for an equation editor I think we need to be able to
>recursively imbed fields within fields and to have the container fields
>expand in size to accomodate them. Is this part of the design?
Actually, fields should probably be an orthogonal issue here. My belief is
that fields will just be a container for "normal" formatted text, with
arguments to allow those contents to be recalculated as needed.
In contrast, I've been expecting that as far as the rest of AbiWord is
concerned, the entire equation is a black box of known size -- kind of like
images, actually. The formatter just treats it as a single unbreakable
cell, and all of the additional calculations and manipulations would get
isolated inside the equation editor / renderer component. Even if this
isn't a reasonable assumption, I still expect that any math-breaking logic
(which would create two formula cells in a row) should be isolated to the
equation editor, thus keeping the rest of the formatter (blissfully)
ignorant.
Have you decided what style of markup you'll want to use for the equations:
- MathML (a la Mozilla or Amaya),
- TeX/LaTeX (a la LyX), or
- something else entirely?
If you're planning to go with MathML, then we'd probably want to upgrade to
expat 1.1 so that we could take advantage of its new namespace support.
That way, the contents of the <math> </math> container would all get managed
by the equation editor -- either as part of the piece table, or perhaps in
its own undo-savvy data structures.
On the other hand, it's less obvious to me how you'd intermingle XML and TeX
syntaxes in an unambiguous way, but I assume there are ways to make the
parsing problem tractable.
Does that sound at all reasonable? I hope so, but I have to admit that
we're really stretching the state of my ability to help frame the problem
here.
Paul
PS: I can't *tell* you how glad I am to not have to be implementing any of
this code. Getting equations to format nicely from an interactive GUI
presents all sorts of usability challenges, so I'm thrilled to have a group
of finicky expert users taking charge of this stuff.
This archive was generated by hypermail 2b25 : Tue Feb 29 2000 - 19:05:49 CST