eml-project changes needed
Changes as decided upon at the Sevilleta EML meeting, April 24-25, 2002:
1) make personnel an element that extends ResponsibleParty with a "role" element
2) make siteDescription and designDescription local elements in the schema
instead of global as they are now
3) rename siteDescription to "studyAreaDescription" -- expect it to evolve to
follow siteDB in later versions. James, David, Karen to take lead on this.
3) revise sampling to include:
studyExtent -- coverage and /or a paragraph*
samplingDescription (paragraph+ incudes description of stratification)
spatialSamplingUnits? -- choice of (geographicCoverage+) or (a reference to
an entityID) -- I don't really understand how this will work, so be sure to
document the semantics fully.
4) link to coverage instead of the individual temporalCov and GeoCov elements
#3 Updated by Matt Jones about 18 years ago
Thanks. Personally, I like the approach 1 better because of its simplicity, but
Corinna's point is right that the appraoch 2 style where users are prompted for
information might get more complete metadata. However, if approach 2 is not a
comprehensive set of site descriptors, then people might be frustrated. You
really need to discuss this with James and Karen and propose a concensus opinion
that you think will work for all ecologists (marine ecology, oceanography,
That said, I see a number of problems with the current drafts you sent that will
need to be fixed before we can check them in. Here goes...
1) There is an error in how you tried to import geographic coverage. You need
to create an element named "geographicCoverage" (fully spelled out) and give it
the type "cov:GeographicCoverage". Likewise, you imported coverage wrong, in
that you used the cov:coverage element rather than creating your own "coverage"
element and assigning it the type "cov:Coverage". We agreed to do imports like
2) You are missing documentation for many tags, and I think for all of the
elements you created. You need to provide full documentation in the form of a
doc:tooltip, doc:summary, doc:description, and doc:example, and doc:lineage for
every element and attribute in the schema. These should fully explain and
define the semantics of each element and attribute in language that is fully
understandable by ecological scientists.
3) Indentation is messed up. Our standard is to use 2 space indents and a
maximum line width of 80 characters for all schemas, including documentation.
You edited with XML Spy, and it raised hell with the existing format. Please be
sure that you have properly formatted the document. Also, note that it would be
better if you made NO changes to whitespace on lines that you are not changing
content -- otherwise, it is more difficult for us to see what you changed using
4) personell needs to be repeatable
5) studyExtent/coverage is not imported/referenced/typed correctly
6) citation is imported incorrectly in all cases. Change to the format
described in (1)
7) samplingDescription/paragraph should be required so that people have to have
at least one paragraph. same with other paragraphs, butonly if other content
8) protocol is imported incorrectly
9) I still don't understand how spatialSamplingUnits works. Maybe when you
document the elements I will understand.
10) The root element should be "researchProject". We agreed to get rid of the
"eml-whatever" root elements in favor of camel caps. Identifier needs to be the
first child of researchProject. There needs to be a complex type called
"ResearchProject" that has the content model of the current "researchProject"
element, and then the "researchProject" element should be changed to be of type
"ResearchProject". This will be used in importing in other locations.
11) I don't understand the difference between "classificatinSystem" and
"descriptorThemeSystem". Again, documentation would probably help.
That's about it. When you make the changes, you can obsolete the attachments
above with new attachments so everyone can see them.
David Blankman wrote:
I have eml-constraint finished except for some documentation. I have
working with Corinna, but we both had some questions about the meaning
of one of the elements. I was unable to clarify with Peter today, but
will do so tomorrow. As soon as I have the clarification from Peter, I
will send you off a patch.
I have two prototypes for eml-project (assigned to James, but delegated
to me). James has been out-of-town but should be back tomorrow.
I am attaching both prototypes. James has not seen either of them.
Prototype 1 (eml-projectR1.xsd) is simple and flexible, it allows for
the use of any one or more classification systems, e.g. bailey, biome,
and allows for any arbitrary set of themeSystems, e.g. climate,
The second prototype (eml-projectR2.xsd) has bailey and biome
classifications as specific elements and the simple/flexible model from
Prototype 1. The second prototype has specific elements for climate,
soils, geology, hydrology, and disturbance themes as well as the
simple/flexible model from Prototype 1.
#5 Updated by David Blankman about 18 years ago
Simplified structure of studyAreaDescription significantly. Deleted classification system. Classification systems will be described as descriptors.
Descriptors now have an attribute of name. Created a DescriptorNameType (modeled on RoleType in eml-party.xsd -- union of enumerated list and unrestricted string) The enumerated list currently includes: climate, soils, geology, hydrology, soils, distrurbances, bailey,biome. There are probably other general site descriptors themes that could be added to the suggested list. The list clearly has a terrestrial bias.
Descriptors have values. The value element has an attribute of "name_or_id". For a published system like "Bailey" there are IDs/codes like M131 and names "open-woodland-tundra" M131 would correspond to the "name_or_id" while "open-woodland-tundra" would be the value. In the case of a "theme" like climate. The "name_or_id" might be "average rainfall" and the value would be "12.5 inches".
Examples of some of the more complex areas such as "samplingDescription" have been left blank. If desired, someone with content expertise should provide the examples.