Overloading Semantics For SBML Objects By Type
Overloading semantics for SBML objects by type
| Object type | Overloadable | Convergeant Overloadable | Divergent Overloadable | Semantics | Comments |
|---|---|---|---|---|---|
| species | yes | no | no | flattened species has union of the species references incident to all the relevant species | |
| reaction | yes | yes | yes | flattened reaction has only the species references incident to the overloading reaction | |
| compartment | yes | no | no | ||
| species type | no | no | no | should use inclusion instead - covers all elements taht are used to assemble a complex type | |
| compartment type | no | no | no | should use inclusion instead | |
| rule | yes | no | no | not clear how you will refer to a rule as they don't have an identity seperate from variable assigned | |
| constraint | no | no | no | really don't want this to be overloadable as constraint will tell the user that the submodel is incorrect | |
| event | yes | no | no | same issues as rule with identity | |
| link | no | no | no | ||
| instance | yes | no | no | the set of ports must match | not really clear what happens when ports are bypassed |
| port | yes | no | no | linking constraints from this table are taken from original object (speciesPort acts as a species) |
-- Andrew Finney, 10 Sep 2007
(Comment from Magdeburg): Convergent links should not be allowed in general. To allow for m to n m replacement of reactions, we propose to define a set of reactions and to link from this set to the reaction to be replaced.