Hi,
I believe the document is in general pretty good. These are my comments to further polish it.

page 4, "Basic Ideas": I think the discussion should be a bit more general here. The reader cannot understand what 
problems the new protocol addresses.
Menu item 1: I think one should spend some time  justifying the introduction of READY message (p6 "Opening Data Channel"
: avoids the use channel count), "CLOSE" (helps controlling bandwidth utilization), etc.
Menu item 2: The same is true here: why the protocol does not use EODC (explained at p14: in order to distinguish 
between EOF and connection dropped).

page 4, "Data Streaming": it is intuitively clear that using transactions ID allows transferring multiple files over the
 same channel. You may want to make the algorithm more explicit.

page 7, "Data Channel Closed by the Sender", I think the sentence "Failure to send 'BYE' message..." is confusing. The 
reader may not understand that the sender *must* wait for BYE from the receiver, as written at p12 1st line (if BYE were
 not waited for, the receiver could try to send a "resend" while the sender is dropping the connection: the receiver 
would not know whether the sender received the "resend" or not). I would rephrase saying that despite the fact the the 
sender must wait for BYE, if for some reason the connection drops before the BYE is sent AND the receiver has a non 
corrupted file, THEN the receiver should not fail.

page 7, Fig 3.: the labels EOD and BYE are misplaced.

page 10, "Active Sender": Sender must not send EOF on any channel until it receives "READY" on all open channels. You 
may want to expand on what it can go wrong if the protocol would not do it.

page 13, 1st menu item of the page (cksum): Parameter value is the keyword specifying the algorithm (reference p16 where
 there are recommended keywords). There's also a typo (first word of the line): is --> if.

page 14, "Explicit EOF Communication in Stream mode", 2 paragraph: OPTS is mentioned without defining it. A reference to
 p16 or a brief description would be nice

General comments:
- It would be useful to have an introductory paragraph with the intended audience.
- You may want to review the use of articles in the text.

Gabriele Garzoglio
SAM-Grid team
DZero-PPDG