Re: [Cdk-devel] atom type mapping

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

On Wed, Jun 11, 2008 at 2:16 PM, Rajarshi Guha <rg...@in...> wrote:
> Hmm - is there a specific advanatage to this? Why not just have separate
> mapping files?

The are... it's not by default using something like cdkat:subTypeOf or sameAs...

> Also, what if you wanted to map from Jmol to MMFF94? I assume
> it go as Jmol -> CDK -> MMFF94?

Yes, that's what I had in mind...

>> One problem concerning mapping is that mapping itself... 1:n, n:1, 1:1
>> etc...
>
> I think it depends on which way you're going. The examples you provide below
> are correct. But what about going from Jmol to CDK? we'd need to make some
> arbitrary decisions on which CDK atom type would map to the Jmol C type.
> This would obviously affect subseuquent conversion to other atom types

Insufficient info I'd say... a jmol:C can be any of the cdk C atom
types... so, such a mapping would not be possible...

Egon

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