Everyone using Daffodil with XML wants to use a variety of XML Schema-oriented tools with that XML.
Unfortunately, experience has taught us that the include/import behavior for schemaLocation varies across tools. The XSD standard from w3c really only says schemaLocations are "hints", resulting in very different interpretations of how they should work.
Only some XML tools allow plugable schema location resolvers, and only Java based such tools can plug in the Daffodil resolver.
Experiments have shown that the only common denominator that works on every XSD tool is all files of the schema live in a single directory so that import/include is always able to find a "peer" file.
So a feature to "flatten" the schema's tree-like directory structure into a single directory is needed so as to provide a version of the schema that can be used by XSD validators. This must do renaming of all the files to avoid collisions and then adjust the include/import statements accordingly. It also must grab files from all the components that make up the schema, as many schemas have multiple components now. It also must cope with the includes of Daffodil-supplied include files.
Everyone using Daffodil with XML wants to use a variety of XML Schema-oriented tools with that XML.
Unfortunately, experience has taught us that the include/import behavior for schemaLocation varies across tools. The XSD standard from w3c really only says schemaLocations are "hints", resulting in very different interpretations of how they should work.
Only some XML tools allow plugable schema location resolvers, and only Java based such tools can plug in the Daffodil resolver.
Experiments have shown that the only common denominator that works on every XSD tool is all files of the schema live in a single directory so that import/include is always able to find a "peer" file.
So a feature to "flatten" the schema's tree-like directory structure into a single directory is needed so as to provide a version of the schema that can be used by XSD validators. This must do renaming of all the files to avoid collisions and then adjust the include/import statements accordingly. It also must grab files from all the components that make up the schema, as many schemas have multiple components now. It also must cope with the includes of Daffodil-supplied include files.