Re: [Mtt-developers] Re: Legacy examples & components.

[Geraint <geraint geraintpaulbevan freeserve co uk>]

Lars Clausen <lrclause cs uiuc edu> writes:

> On Thu, 28 Nov 2002, Peter Gawthrop wrote:
> > Hi Dave & Geraint,
> >
> > just to say that it is important to retain fig for the time being, and
> > indeed allow systems with components with either fig and dia
> > representation -- I don't see a problem with this, do you?
>
> We have no intention of removing the fig import and export components from
> Dia, but there are things in Dia that cannot be expressed in Fig and vice
> versa, so it cannot be an underlying file format.
>
> -Lars

Hi Lars,

as things stand at the moment, MTT works with bond graphs built with Xfig and looks to be well on the way to 
working with bond graphs built with Dia. The way that MTT works is to transform models between various 
representations, often for use with different tools or languages, by way of other intermediate 
representations.

I don't think that we will have any problem adapting MTT to work equally well with models built with either 
of the drawing packages - we just need to define a route to a common representation from each format. This 
common representation (or set of representations) must contain the information necessary to generate 
equations, but it does not have to contain the graphical information. Therefore, I don't think that 
importing/exporting drawings between Xfig and Dia will really be an issue for us. The problem that we face is 
getting MTT to work with models created with one tool that contain sub-models built with the other.

As bond graphs are hierarchical, it will be necessary for us to be able to generate equations from models 
such as these, however I don't expect that we will attempt to do this by converting (or importing) models 
from one format to the other, or by defining a common graphical format. Rather, we should be able to 
accomplish our goal by ensuring that MTT can work at each level of the hierarchy in whichever format the 
model happens to have been drawn in.

By the way, thank you for being so receptive to Dave's suggestions for improvements to Dia to help with the 
integration with MTT.

/Geraint.