Morpho

the map.remove() should be part of the setPanelData() method for the
enumeratedDomain subpanel, not the setPageData() method for the top level
attributePage. That way, it is located logically where the relevant metadata
would be handled (or will be handled if, in future, a checkbox is ever added
for this attribute, as chris mentioned above.)

As an addition to this bug, some elements may have been overlooked as well:

/eml:eml/dataset/dataTable/attributeList/attribute/missingValueCode/codeExplanation

is left over in the OrderedMap, and so the Tree Editor is enabled rather than
the Attribute Editor

One more attribute not handled and left in the OrderedMap:

/eml:eml/dataset/dataTable/attributeList/attribute/storageType/@typeSystem

An alternative 'fix' would be to change the way the tree editor handles xml
attributes. Right now it inserts suggested value wheng available and trims those
with no information when saving. The template could be changed to not insert
anything into all optional attributes. Unfortunaely, there is currently a bug in
the tree editor in that it does not 'know' whether xml attributes are optional
or required.

I've changed both AttributePage.java and NominalOrdinalPanel.java to handle the
/@id, /@scope, /@system, /@enforced, /@typeSystem attributes. We still need to
address the missingValueCode/codeExplanation element mention previously. Also,
we need to check both IntervalRatioPanel.java and DateTimePanel.java for similar
bugs where there are not placeholders or widgets to handle the various XML
attributes.

I've added placeholders for the <missingValueCode> sub elements:
and <codeExplanation>. This changed the setPageData() and getPageData methods.

Trying to summarize the state of this bug based on above comments and cvs
messages for commits done by Chris.
(http://cvs.ecoinformatics.org/cvs/cvsweb.cgi/morpho/src/edu/ucsb/nceas/morpho/p
lugins/datapackagewizard/pages/AttributePage.java
and
http://cvs.ecoinformatics.org/cvs/cvsweb.cgi/morpho/src/edu/ucsb/nceas/morpho/pl
ugins/datapackagewizard/pages/NominalOrdinalPanel.java)

1. "This fix brings up another bug: <enumeratedDomain> is repeatable, but the
AttributePage class is only set up to handle the first instance of this element
by using the NominalOrdinalPanel, IntervalRatioPanel, etc. "
This should be added as a bug if not added already

2. "Also, we need to check both IntervalRatioPanel.java and DateTimePanel.java
for similar bugs where there are not placeholders or widgets to handle the
various XML attributes."
Will add placeholders for @id to IntervalRatioPanel.java and
DateTimePanel.java

3. "Still need to deal with coverage and method. This fix also only supports a
single missing value code/codeExplanation pair for the attribute in question.
When this is converted to a widget in the interface, we should handle the case
were missingValueCode elements repeat."
We do not cover a lot of the eml fields in the wizard screens. This is true
for all the wizard screens. This is because wizard screens were kept simple and
only used for gathering data which is required. Hence, though having widgets
for coverage and method information would be useful, there might be some
discussion on which fields should be included in the screens. Same is true for
missingValueCode.

I think placeholders for attributes like id,scope etc are good to have. But any
field that has relevant data in it should be displayed to the user. Thats where
tree editor comes in - even though it is not as user friendly as wizard
screens.

Original Bugzilla ID was 1561