Some of it seems, to my eyes, to be uninformed by the community that works in the AQM space, so I see perfectly serious quotes like

Longer packets might be dropped in favor of shorter ones. Packets in a
burst might be dropped in favor of ones that are spaced out.

This gets back to the point of neutrality - which describes some level
of "equivalence", but there's never just one version of equivalence that
everyone will accept.

If you want to preserve the Internet architecture, you need to make sure
that:

Packets shall not be discriminated except on their inherent
properties (size, time of arrival) or explicit user-inserted
label (e.g., a QoS tag).

If you want to make sure that packets are "fairly dropped", there's no
single such thing; one link might be bandwidth limited (so drop
proportional to length is fair), and the next might be header processing
limited (so per-packet drop is fair); for a given path, there's no
single mechanism that satisfies the variety of fairnesses that could be
required.

If anyone's also interested in Network Neutrality and its roots, please feel free to hop over and contribute some informed opinions (;-))

--dave

-- 
David Collier-Brown,         | Always do right. This will gratify
System Programmer and Author | some people and astonish the rest
davecb@spamcop.net           |                      -- Mark Twain