text import issues with eml-attribute creation
While parsing and importing text files we currently default fill in the
attribute definition to be identical to the attribute name. This is wrong.
Instead, we should have separate boxes for the attribute name, label,
definition, and units as part of the text import dialog. As the definitin is a
required field, the import wizard should not let the user move on until all
fields are defined (or post a warning like it does for other content violations).
#1 Updated by Matt Jones over 18 years ago
This is partially completed. However, when I tested it in Morpho the values
that I entered into the fields for label, definition, and units were not
properly saved. In addition, the name field requires the user to hit return in
an unnatural way in order for the value to persist. Overall, need to
systematically evaluate the usability of this form field.
#3 Updated by Matt Jones over 18 years ago
Better. Now the relevant fields are present and they save properly. Some new
issues have arisen.
1) The "unit" field is required to move on, but is not required in eml. Need to
only require those fileds that eml requires.
2) When outputting a text column with the "Include in metadata" box checked,
some spurious values are put into the rangeDomain element, when in fact
rangeDomain should be omitted altogether for text type columns.
3) enumeratedDomain should be excluded for numeric columns if rangeDomain is
4) The code definitions in enumeratedDomain are blank, empty elements. The
wizard should prompt the user to defineeach code or provide a source if they
choose to include the enumerated domain in the output. In no case should empty
elements be output.