Some ideas about multiple source levels. They are a commons problem in my research and I have found no program that can represent them correctly.
an exmple:
Sources should be first-class resources, just like an Individual, an Event, a Place, or a Date. For example, here's how you might show that the information in the <#data> graph came from a book:
<#data> gp:source <#book> . <#book> dc:title "Title of the Book" .
To specify the page number that the data came from, you might instead do this:
<#data> gp:source <#page> .
<#page>
gp:page 236 ;
gp:part <#book> .
<#book> dc:title "Title of the Book" .
The alternative is to relate the graph to both a page number and a book, like so:
<#data>
dc:title "Title of the Book" ;
gp:page 236 .
This is undesirable because when you are working with a set of graphs obtained from different people, one person may claim that a piece of data came from one source, while another may claim that it came from a different source. The merge of this data can be confusing:
<#data>
dc:title "Title of the Book" ;
dc:title "Title of a More Different Book" ;
gp:page 117 ;
gp:page 236 .
Whereas, with first-class sources, you would get the following, which allows the interpreter to distinguish clearly between sources:
<#data>
gp:source <#page> ;
gp:source <#page2> .
<#page>
gp:page 236 ;
gp:part <#book> .
<#book> dc:title "Title of the Book" .
<#page2>
gp:page 117 ;
gp:part <#book2> .
<#book2> dc:title "Title of a More Different Book" .