I'm documenting a source in gdmxml now, and as I expected a lot of
questions have been raised in my mind. 

First is the issue of citation-parts.  The citation-part is used to
store citation parts such as title, author, page, etc. As it stands now
I would document "Marriages in Fladstrand Parish (Film #0049002), p.
199)" [1] as

  <source id="1" />
  <citation-part source="1" citation-part-type="1">
    Marriages in Fladstrand Parish
  </citation-part>
  <citation-part source="1" citation-part-type="2">0049002</citation-part>
  <citation-part source="1" citation-part-type="3">199</citation-part>
  <citation-part-type id="1">Title</citation-part-type>
  <citation-part-type id="2">Film #</citation-part-type>
  <citation-part-type id="3">p.</citation-part-type>

But it seems natural to put the citation-part elements inside a source
element, like this

  <source id="1">
    <citation-part citation-part-type="1">
      Marriages in Fladstrand Parish
    </citation-part>
    <citation-part citation-part-type="2">0049002</citation-part>
    <citation-part citation-part-type="3">199</citation-part>
  </source>
  <citation-part-type id="1">Title</citation-part-type>
  <citation-part-type id="2">Film #</citation-part-type>
  <citation-part-type id="3">p.</citation-part-type>

That's easy enough to specify, and it makes sense because there's one
and only one source per citation-part.  Which brings us to
source-groups. These two relationships hold: One source-group has zero
to many sources, and one source belongs to zero to many source-groups.
In a database we definitely want a source-group-source table to get rid
of this many-to-many relationship. But again we can collapse things a
bit:

  <source id="1">
    <source-group-source source-group="1"/>
    <source-group-source source-group="2"/>
  </source>
  <source-group id="1">group 1</source-group>
  <source-group id="2">group 2</source-group>

Does this approach seem valid? (This would apply in many other places
throughout the spec.)

You'll probably also noticed I have decided to leave off the "-id" on
the end of IDREF attributes. It just feels redundant to have it there.

Also, in my hand-cooked examples I used the ID "1" (for example) more
than once, but for different attributes.  Is that valid, or do IDs need
to be unique even when identifying different element types?

Hans :)

1. http://fugal.net/andreas/first_fugl/fugl1.htm footnote 2.
-- 
"Everybody is talking about the weather but nobody does anything about it."
        -- Mark Twain

_______________________________________________
gdmxml mailing list
[EMAIL PROTECTED]
http://fugal.net/cgi-bin/mailman/listinfo/gdmxml

Reply via email to