From: Jones, A. <And...@li...> - 2008-06-06 14:56:30
|
Hi all, I've made a few edits to the schema as discussed on the call, and tidied up the SVN, so the main working version is the only file in the schema directory. Changes: - Altered cardinality of AnalysisCollection and AnalysisProtocolCollection to 1, since I think it should not be possible to create a file without any protocols or protocol applications (as least if there is a protocol application there must be a protocol). Let me know if anyone disagrees and I'll change it back - Removed PolypeptideValuePropertyType and ValuePropertyType, Intensity, RetentionTime (think these were left over in the schema and not required) - Put the AnalysisXML complexType first in the schema to make it easier to find. - Worked thru schema and put in PropertyValue (value + unit + unitAccession) references instead of cvParam where sensible... - Added the option for userParam or cvParam by adding an abstract ParamType (changed over most of the references to ParamType, although I may have missed a few) - Made TranslationTable cardinality 1, as suggested by Martin’s email One minor problem. On the call we discussed using XSD style inheritance for cvParam, where an entity e.g. SearchType uses the cvParamType complexType. Unfortunately, this does not work with also using the other style of inheritance where you can give the option of a userParam or cvParam by giving a reference to an abstract superclass e.g. ParamType. So... we either go back to the old style of having e.g. <SearchType> <cvParam ...> Or we ditch the option of having cvParam or userParam as an option, and only user cvParam. We could for example make cvRef and accession optional, so a userParam is just a cvParam without the reference. I can give examples if I haven't explained this clearly enough. Cheers Andy |