Technical remarks
      
In this paper the main result of Naor, et al.[1] and its proof is being
presented. The main result of [1] is a construction of an uncoditionally 
secure message authentication protocol using manually authenticated channel
such that the length of the manually transmitted string is about 2 log 1/e, 
where e is the attack probability. 

The paper is not easy and Vesa has done a good job reading and working through
the proof. He has also understood all steps.

The paper could be improved by giving a survey of other related results
[4,5]and what is the significance of this particular result in the general 
context. 

Also adding a concrete example of typical parameters would help the reader 
to understand the construction. Below is one typical set of parameters:

k = 3
e = 2^{-12}
n_1 = 2^{18} (that is, 32 kB)
Q_1 = 2^{32}
n_2 = 64 = 2^6
Q_2 = 2^{19}

This also demonstrates the fact that in this construction, the length of the 
message has only a small effect to the length of the manually transmitted 
string, which in this case is 38 bits,  while \epsilon is the most dominant 
factor. See example parameers for a very short message:
 
k = 3
e = 2^{-12}
n_1 = 2^{2} (that is, 4 bits)
Q_1 = 2^{16}
n_2 = 16 = 2^4
Q_2 = 2^{17}
   
(Please check)

      
The proof follows the one presented in the original paper. There are no novel 
techniques.


Editorial quality

Vesa's paper is quite easy to follow. It follows quite faithfully the 
presentation in the original paper, sometimes even sentence by sentence.

Typo's, other errors and suggestions for improvement:

page2, column1, line2: parsed as

p.2, c.1, l.7-11: This may be too trivial to mention here

p.2, c.1, l.14: points

p.2, c.1, l.23 and 24: \leq (less than or egual to) k  [the highest degree terms may cancel]

Section 2, paragraphs 1-2: the parameter k here is not related to the number 
of rounds of the protocol denoted also by k later on.

p.2, c.1, l.34: \epsilon [not e]

p.2, c.1, l.36: GF[Q_j]

p.3, c.2, l. -13-12: then to know whether adversary chose 

(I stop here as the rest of the paper is still subject to change)