At 10:41 13/12/2002 +0100, Konrad Hinsen wrote:
Gustavo Mercier <gamerc...@yahoo.com> writes:
>As a suggestion, let's consider the chemistry implementation of XML -- CML. Whatever the details of the object "molecule", it >would be beneficial to input and output to/from CML. A specification of "molecule" based on CML would make it easier to interface >with other technologies. I recognize that there are problems with CML, but this should not stop the community from considering >this "standard".
>Gustavo
> As a suggestion, let's consider the chemistry implementation of XML > -- CML. Whatever the details of the object "molecule", it would be > beneficial to input and output to/from CML. A specification of I don't agree. I looked at CML a while ago, thought about implementing it in MMTK, but quickly found out that almost none of the information I needed to store about molecules could be represented in CML, except by adding "conventions" of my own - but then I could just as well define my own XML format. CML is based on the right intentions, but is too weak as an implementation.
CML is designed to be extensible by a variety of mechanisms such as external ontologies and we would be delighted to hear from those who find it "too weak". We are currently developing a computational chemistry implementation (CCML) as previously mentioned on this list. The feature of XML languages is that XML requires semantics to be well designed and interoperable, *and* to have communally available software tools for reading, validating and writing and linking to ontologies. CML is designed for collaborative working and we shall shortly announce the mailing list.
P. Unilever Centre for Molecular Informatics Cambridge University, UK
