Skip to content

Guidelines for choosing data names #10

@jamesrhester

Description

@jamesrhester

There are a number of non-obvious rules for choosing data names, which we should document. The ones that spring to mind are:

  1. A data name should be formed as _<category>.<object>.
  2. Characters in data names are restricted to alphanumeric symbols, _ and .
  3. An object data name section = special_details is usually used to record information not captured in other tags of the category
  4. _su should be appended to object to represent the SU of Measurand data names
  5. Where category B is a child category of category A, all <object> values of B should be distinct from any <object> values in A, with the exception of linked key data names. This is because dREL may refer to items in B as if they are part of A using only their object_id.
  6. Where Loop category B is a child category of Loop category A, the category name for B would normally be prefixed with the category name for A, followed by an underscore. For example, atom_type_scat is a child category of atom_type.
  7. Where a category provides overall information for looped items in a separate category, conventionally s is appended to the looped categories name, for example, reflns provides summary information for reflections in the refln loop.
  8. When creating DDLm data names to match legacy DDL1 data names, minimal changes to the ideal _<category>.<object> form should be made
  9. Not all data names conform to these rules due to the need to consider legacy naming from DDL1 dictionaries.

If these sound like a good start, we should create a document.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions