Overall
-------

From tht title it sounds like your focus is on the use of out-of-band
channels (which you call "physical contact") in Secure First
Connect. But your survey is very incomplete: quoting from the seminar
website:
 "You need not (and should not) limit yourself to this starter
 material only."

Try making an exhaustive survey of first connect mechanisms using
different out of band technologies comparing the OOB technologies
in terms of their security properties.  In parallel, you can describe the
actual key agreement protocols in greater detail and analyze them, focussing
on what kind of security requirements they impose, when they are used with
OOB technologies.  As discussed in the first meeting, you should take Jukka Valkonen's
thesis as the starting point.

Technical
---------
 
Section 2
- "transient association" is not a necessary assumption
- "no previous context" and "no trusted third party" are two aspects
  of the same thing: initialization of security association has to be
  done by the users themselves.
- the constraints you quote from Jan-Erik's paper are valid, but you
  don't use or refer to these constraints afterwards.

Section 4
- the barcodes or the blinks used in the visual channel are not
  "secret".

Section 5
- restricting that each device must run only one pairing protocol
  instantace at a time is probably a reasonable restriction in
  practice.   You could explain in more detail what kind of security
  problems you see in allowing parallel runs of pairing protocols.

Editorial
---------
- Explain acronyms on first use (e.g. PSP=
- be consistent in references
  e.g., ref 4 lists the first name of the first author; ref 6 only
  lists initials.