John,
I'm fine with the renaming, but could we eventually make it bit easier in
1.6 by
1) Make Molecule class just plain descendant of AtomContainer, without any
differences between classes.
2) Mark AtomContainer deprecated.
3) At a later point of time remove Molecule and rename AtomContainer to
Molecule, hence AtomContainer should disappear, but after being deprecated
for a while.
Best regards,
Nina
On 4 October 2013 17:32, John May <[email protected]> wrote:
> *Too Long Didn't Read:*
> * There is no more Molecule in new CDK versions only AtomContainer, this
> could be confusing - vote here: http://strawpoll.me/501772*
>
> Hi all,
>
> As you may (or may not) know the CDK main chemical structure object is the
> *AtomContainer* and *Molecule* (excluding protein structures polymers
> etc).
>
> In the *CDK 1.4 *there were both instances and the distinction was a *
> Molecule* is connected (but this was not checked). If you are using *CDK
> 1.4 *code looks something like this.
>
> MDLV2000Reader reader = new MDLV2000Reader();
> IMolecule mol = reader.read(new Molecule());
>
>
> The new version of *CDK. 1.6 *will have API changes one of which is there
> is no more IMolecule/Molecule.
> This simplifies interactions with the rest of the API as you no longer
> have to create a new Molecule() when you have an AtomContainer however
> it does mean the main chemical structure class is now an *AtomContainer*.
>
> MDLV2000Reader reader = new MDLV2000Reader();
> IAtomContainer container = reader.read(new AtomContainer());
>
>
> This is *semantically correct* for the current inheritance hierarchy with
> *Crystal* extends *AtomContainer *etc. However I believe that the name *
> AtomContainer* is not
> intuitive and having discussed with many people this seems to be the
> consensus. New users I've helped find it particularly confusing. I
> understand the semantics
> argument but I strongly believe the object model should not try to exactly
> represent reality. In particularly we should consider the domain of
> applicability and ease of use.
>
> I propose we rename *AtomContainer* to *Molecule - *this is a huge API
> change but remember if you will be upgrading from *1.4* versions you will
> have to rename
> *Molecule* from you code already - with this change you would have to
> rename *AtomContainer* instead.
> API changes are a pain but as the toolkit has evolved there are inevitable
> incompatibilities. When the release happens we will provide some scripts to
> patch you code.
>
> I have discussed with Egon and we decided this of course needs input from
> the mailing list. If you have a concise suggestion for a name we are also
> happy to hear it.
>
> Most importantly please vote for you preferred choice here:
> http://strawpoll.me/501772
>
> Some names from other toolkits
>
>
> - RDKit: RWMol,
> ROMol<https://github.com/rdkit/rdkit/blob/master/Code/GraphMol/ROMol.h>
>
> - ChemAxon: Molecule (extends MolGraph)
> - Open Babel: OBMol
> - PerlMol:
> Chemistry::Mol<http://search.cpan.org/dist/Chemistry-Mol/Mol.pm>
> - OUCH:
> Molecule<https://github.com/odj/Ouch/blob/master/Ouch/Structure/Molecule.hs>
> - Daylight:
> molecule<http://daylight.com/dayhtml/doc/man/man3ob/molecule.html>
>
> Many thanks,
> John
>
>
> ------------------------------------------------------------------------------
> October Webinars: Code for Performance
> Free Intel webinars can help you accelerate application performance.
> Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most
> from
> the latest Intel processors and coprocessors. See abstracts and register >
> http://pubads.g.doubleclick.net/gampad/clk?id=60134791&iu=/4140/ostg.clktrk
> _______________________________________________
> Cdk-devel mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/cdk-devel
>
>
------------------------------------------------------------------------------
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from
the latest Intel processors and coprocessors. See abstracts and register >
http://pubads.g.doubleclick.net/gampad/clk?id=60134791&iu=/4140/ostg.clktrk
_______________________________________________
Cdk-user mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/cdk-user
