On Nov 10, 8:50am, Jan Zelinka wrote: > Subject: Re: Draft DDL Version 2.0.8 Released > Dear John, > > I had a brief look at the previous version. It looks fine to me (although a bit > overcomplicated), but could you save my time and give me an example, how do > you map parent/child relationship on the relational schema? > > Why do we need both? I do not understand the semantics... > > Cheers > > J. > > > >-- End of excerpt from Jan Zelinka Hi Jan, Your point is well taken, the parent-child relationships are over specified. In the original rddl, only the parent information was specified when an item was defined. In the current ddl, all instances of a data item are defined in the parent category. So in the ITEM category there must be a declaration for each data item which names the item and identifies its category. In the ITEM_LINKED category, you then enumerate each child instance of the item ( the parent item can be implicitly derived from context). This organization was used in order to accomodate the CIF/MIF need for explicit declaration of the child item names. Most of the changes that have been introduced into this version of the DDL have been for the purpose of providing some mechanism to map back to what is now done in CIF and MIF. The use of the full data item names as the keys for the many of the DDL categories is certainly not pretty from a relational perspective, but it does provide a bridge to small molecule DDL. The best illustration of the relational schema of DDL is in the diagrams in the introductory document on the server. Regards... John -- **************************************************************************** * John Westbrook Ph: (908) 445-5156 * * Department of Chemistry Fax: (908) 445-5958 * * Rutgers University * * PO Box 939 e-mail: jwest@rutchem.rutgers.edu * * Piscataway, NJ 08855-0939 * ****************************************************************************