Forums

I look at your picture and wondered why you need to visually indicate the center segment? Also why have a separate central segment? Why not join the arcs directly? Is there some biological significance to this?

Herbert

-----Original Message-----
From: Sven Sahle [mailto:sven.sahle@eml.villa-bosch.de]
Sent: Mon 7/14/2003 3:03 AM
To: Herbert Sauro; sbml-discuss@its.caltech.edu
Cc:
Subject: Re: [sbml-discuss] Layout Extension Update



Hello,

it seems that, again, we will have to make some general decision:
Do we want to have several ways to represent reactions in our layout (to
reflect the different way existing programs deal with reactions) or do we
want to find one way to represent reactions.
The one way would need to be fairly general to allow most other cases to be
translated. It would therefore be unnecessarily complex for some simple
cases.

In our proposal we tried to find a compromise between generality and
simplicity:
A graphical representation of a reaction in our proposal consists of the
coordinates of exactly two nodes. One node is where the substrate arcs meet,
the other is where the product arcs meet. Optionally information about the
curve connecting the nodes could be given (not yet in our proposal). Also the
graphical representation of the reaction contains a list of graphical
representations of species references. These contain information on which
species glyphs connect to which node of the graphical representation of the
reaction. Optionally also information on the form of the curve.
(In the attached figure the red squares are the nodes of the graphical
representaion of the reaction. The green curves are the graphical
representations of the species references. The blue circles are the graphical
representations of species. So the red and green objects together form the
graphical repsresentation of the reaction)

Of course more general graphical representations of reactions could be thought
of (like an arbitrary number of nodes in the reaction).

If a program internally deals with only one node (center of reaction) it could
either give the two nodes the same coordinates or regard them as one point
plus a tangent.

So what do you think? Do we need a more general definition? Do we need several
cases for simple reactions and complicated reactions? Or could there be a
compromise similar to our proposal?

Sven

On Saturday 12 July 2003 21:22, Herbert Sauro wrote:
> You're correct, Jdesigner does have an explict reaction section, instead
> the whole line from reactants to products represents the reaction, thus
> clicking anywhere on the line will bring up the reaction property box. No
> harm though in allowing people to have special symbols or sections of the
> line that represent the reaction.
>
>
> Herbert
>
> >I also agree with Herbert et al about line placement, but we
> >need to think pretty hard about it so that everyone is happy.
> >For example, JDesigner (as I recall) doesn't have a
> >graphical representation for the reaction; instead, the
> >reaction is represented by the collection of lines in
> >an implicit manner. On the other hand, the Virtual Cell
> >group takes an approach similar to the EML approach, with
> >two (x, y) points representing the reaction (one point
> >where reactants hook in, one where products hook in).
> >I think we need to treat both approaches.
> >
> >John