Re: [Cdk-devel] Request for comment: Dative Bonds in JChemPaint and CDK

[Egon W. <ego...@gm...>]

On Dec 17, 2007 11:52 AM, Christoph Steinbeck <er...@do...> wrote:
> we have a user request to add dative bonds to JChemPaint (see attached
> picture).
> The modification requires a new renderer option

The depiction is easy to work out.

> as well as potentially a way to code dative bonds on the data classes level.

The representation in data classes is more tricky...

IBond is the intended class for this, and has a field for electron
count (which is not really used). However, semantics are not worked
out, and existing algorithms will likely assume an IBond to be a
covalent bond... this is also reflected by the current IBond.Order
enumeration, which only contains such bonds.

A new IElectronContainer-derived interface is a solution, but would
imply new and additional API methods for the IAtomContainer... Maybe
ICoordinationBond...

Additionally, I a lot of code needs to be reworked, to define how they
operate on such bonds. Such as the
ConnectivityChecker.partitionIntoMolecules()...

Think of it like this... does it fit the current atom type definitions
used in the CDK atom type list... because that is about how the CDK
works with things... and that has, currently, no room for dative
bonds... the first step in this project, would be to work out those
details...

Egon

-- 
----
http://chem-bla-ics.blogspot.com/