Granularity of Reusable Content

[Scott Chapal]

eml-dev,

The question has come up here during our testing wether it would be
advisable to have other [additional] fine-grained, reusable content
elements, like <phone>, <onlineUrl> or <organizationName>, for
example.  I imagine this idea might to apply to several other elements
throughout EML.

I realize that planning for reuse could be taken to an unwarranted
extreme, but wonder if there is some 'optimal' degree of normalization
for a Schema such as EML?  Could it achieve heightened integrity of
the metadata?  Simplify updates?  Too convoluted to implement in the
schema?

This topic may have been debated during the design of beta9 or
earlier, sorry for the rehash if it has.

-Scott Chapal

mjordan at spock.jonesctr.org writes:

> > Myla,

> >After looking at this for two hours, I don't really understand the
> >dilemma.
> >
> >The way reusability is defined in EML looks to be something of a
> >compromise.  The unit that is extensible (for the purposes of this
> >example) is address.  So when a reusable address is declared within
> >the first declared creator, it is assigned an id="address.1".  Then,
> >when another 'creator' is declared the address is re-used.
> >
> ><phone> and <onlineUrl> are not defined to be reusable, probably
> >because it is assumed that they are tied to an "individual".  
> 
> I understand what you are explaining...I even understand why they
> wrote it the way they did, but in our situation though, the address,
> phone, onlineUrl (if used generically) and even organization name
> can be the same for many (or all) individuals, but you can only
> reuse the address.  I can't imagine we are so different from other
> organizations.  Obviously I can populate the fields for each element
> during the transformation so it isn't a big deal to me - just
> something I noticed.

> >The design of EML may be influenced by [our] feedback: if there are
> >substantive reasons to advocate a finer granularity for reusable
> >Content then we could propose that.  I don't think we have enough
> >experience to do that yet.

> >Comments?
> >
> >-Scott

> >Consider the following valid XML in dataset-eml.xml:
> >
> >  <creator xmlns="" id="creator.1">
> >    <individualName>
> >      <salutation>Mrs.</salutation>
> >      <givenName>Myla Marie</givenName>
> >      <surName>Jordan</surName>
> >    </individualName>
> >    <organizationName>Joseph W. Jones Ecological Research
> >    Center</organizationName>
> >    <positionName>Information Systems Developer</positionName>
> >    <address id="address.1">
> >      <deliveryPoint>Route 2 Box 2324</deliveryPoint>
> >      <city>Newton</city>
> >      <administrativeArea>GA</administrativeArea>
> >      <postalCode>31770-9651</postalCode>
> >      <country>US</country>
> >    </address>
> >    <phone>229-734-4706 x246</phone>
> >    <electronicMailAddress>myla.jordan at jonesctr.org</electronicMailAddress>
> >    
> <onlineUrl>http://www.jonesctr.org/admin/staff/research.support.staff.html</onlineUrl>
> >  </creator>
> >
> >  <creator xmlns="" id="creator.2">
> >    <individualName>
> >      <salutation>Mr.</salutation>
> >      <givenName>Scott</givenName>
> >      <surName>Chapal</surName>
> >    </individualName>
> >    <organizationName>Joseph W. Jones Ecological Research
> >    Center</organizationName>
> >    <positionName>Database and Network Manager</positionName>
> >    <address>
> >    <references>address.1</references></address>
> >      <phone>229-734-4706 x227</phone>
> > <onlineUrl>http://www.jonesctr.org/admin/staff/sechapal.html</onlineUrl>
> ></creator>
> >
> >  <metadataProvider xmlns="">
> >    <references>creator.2</references></metadataProvider>
> >
> >  <associatedParty xmlns="">
> >    <references>creator.1</references>
> >    <role>Metadata Author</role></associatedParty>
> >
> >  <associatedParty xmlns="">
> >    <references>creator.2</references>
> >    <role>Science IT Liason</role></associatedParty>
> >
> >
> >Myla Jordan <mjordan at spock.jonesctr.org> writes:
> >
> >> <creator xmlns="" id="creator.1">
> >>   <organizationName>Joseph W. Jones Ecological Research Center</organizationName>
> >>   <address id="address.1">
> >>     <deliveryPoint>Route 2 Box 2324</deliveryPoint>
> >>     <city>Newton</city>
> >>     <administrativeArea>GA</administrativeArea>
> >>     <postalCode>31770-9651</postalCode>
> >>     <country>US</country>
> >>   </address>
> >>   <phone>229-734-4706</phone>
> >>   <onlineUrl>www.jonesctr.org</onlineUrl>
> >> </creator>
> >> 
> >> <associatedParty>
> >>   <references>creator.1</references>
> >>   <individualName>
> >>     <salutation>Mrs.</salutation>
> >>     <givenName>Myla Marie</givenName>
> >>     <surName>Jordan</surName>
> >>   </individualName>
> >>   <role>Information Systems Developer</role>
> >> </associatedParty>
> >
> >> organizationName, address, phone, onlineUrl is the same for both,
> >> but the individualName would be different - or as in this example
> >> the creator does not have an individualName element, but I want the
> >> associatedParty to have one.
> >
> >
> >> >Myla Jordan <mjordan at spock.jonesctr.org> writes:
> >
> >> >> Can you define other elements when using the references element?
> >> >> For example: If I have a address element defined with id
> >> >> "address.1" within the creator element then can I reference that
> >> >> address id in the references section of another element - say
> >> >> associatedParty element - then also add individualName,
> >> >> organizationName, positionName, phone, electronicMailAddress, and
> >> >> onlineUrl elements?  This doesn't seem to be an option, but would
> >> >> be a useful one.
-- 
\SEC