John, others, On Mon, Oct 7, 2013 at 11:31 AM, John May <john.wilkinson...@gmail.com> wrote: > I'm actually okay with ChemicalEntity but as a more appropriate name for the > existing ChemObject (which AtomContainer already extends).
Now there is a class we must reconsider... ChemObject exists because we decided to have some core properties and no one came up with a better idea; it's the Java approach, I guess. And a Map on each atom and bond is questionable at least... Similarly, we for long want to remove all those *FooSet classes and just List<IAtomContainer>, ummm... List<IMolecule>, ummm.... List<IChemicalEntity>... Rajarshi, now that you have some time, want to push some of your ideas in this area in the master? > I think 1.7+ we'll need a long think about the object hierarchy and tighten > up the definitions a bit. I have no problem representing the complexity of > chemistry Agreed. To me naming of classes is quite orthogonal to the hierarchy... we can simply one, while keeping the other more complex. (where the one is the hierarchy :) > but it shouldn't try to encode an ontology in the classes. We most certainly don't have to replicate a full ontology. I have also the insight that while we should have ontologies to encode isotope information, the running CDK should use a simpler format which is way much faster at loading and smaller in size, and created from the proper source. John, we have a lot to do. But please don't forget, speed never was a goal! Functionality was, and that is what is what made the CDK big. Now we introduce speed, but have to be careful not to loose functionality in the process. We *can* do both. (In fact, a paper a year or two ago showed that the CDK was not really that slow at all... and sometimes even faster than C/C++... the reference should be somewhere in my CiteULike library... still, we can be faster, and you have done amazing work here.) Egon -- Dr E.L. Willighagen Postdoctoral Researcher Department of Bioinformatics - BiGCaT Maastricht University (http://www.bigcat.unimaas.nl/) Homepage: http://egonw.github.com/ LinkedIn: http://se.linkedin.com/in/egonw Blog: http://chem-bla-ics.blogspot.com/ PubList: http://www.citeulike.org/user/egonw/tag/papers ORCID: 0000-0001-7542-0286 ------------------------------------------------------------------------------ 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 Cdk-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/cdk-user
