OK, I'll bite

I'm currently working to encode some of the texts I'm responsible for on the http://nzetc.victoria.ac.nz/ site. These are digitised historical texts published as sets of webpages. I'm currently struggling with the Work / Instance relationship, the issue I'm struggling with is actually duplicated exactly in the Primer at: http://www.loc.gov/bibframe/pdf/marcld-report-11-21-2012.pdf I'll go with the primer example, since it's better known. 

An Instance is defined on page 10 as:

"BIBFRAME Instances reflect an individual, material embodiment of a BIBFRAME Work that 
can be physical or digital in nature. A BIBFRAME Instance exists as a Web based control 
point that includes properties specific to the materialization as well as contextual 
relationships to appropriate BIBFRAME Authorities related to the publication, production, 
distribution of the material resource. Each BIBFRAME Instance is an instance of one and 
only one BIBFRAME Work."

The figure on page 19 has three instances, a phyical, a PDF and an HTML, the HTML being identified as http://archive.ifla.org/VII/s13/frbr/frbr_current_toc.htm 

The conceptualisation issue I have seems to be expressable via the questions:

(a) what is the status in the bibframe model of http://archive.ifla.org/VII/s13/frbr/frbr_current3.htm ?

(b) what is the status in the bibframe model of http://archive.ifla.org/VII/s13/frbr/fig3-1.jpg ?

(c) shouldn't the HTML Instance be defined as the http://bibframe.org/vocab/unionOf all the HTML pages that make up the HTML version of the report plus all their page dependencies (css, images, media files, etc)?

(d) if an instance corresponds to only one Work, in the case of composite Works which are subsequently split into HTML along the lines of the original Works, do the HTML pages correspond to the original or composite Work?

(e) why does the definition from page 10 of the report not appear on http://bibframe.org/vocab/Instance.html Has the definition changed?