Thanks, Mike!

   "PaymentRequest messages larger than 50,000 bytes should be rejected by the merchant's server, to mitigate denial-of-service attacks."

Do you mean "users wallet" here?

Yes, fixed.

 
You could note in the motivation section two more motivations:
1) That the protocol can be a foundation on which other features are built

I don't like putting "this is what we think will happen in the future" types of statements in specifications, so I'm inclined to leave that out.
 
2) That it is required to assist hardware wallets when there is a virus on the system

Added:

"Resistance from man-in-the-middle attacks that replace a merchant's bitcoin address with an attacker's address before a transaction is authorized with a hardware wallet."

Perhaps note in the BIP that the merchant should not assume the merchant_data field is trustworthy - malicious buyers could rewrite it as they see fit. Point out that a good way to use this is to serialize server state, signed by a merchant-only key, in the same way one might use an HTTP cookie.
 
Added:

"Note that malicious clients may modify the merchant_data, so should be authenticated in some way (for example, signed with a merchant-only key)."
 
   "PaymentDetails.payment_url must be secure against man-in-the-middle attacks that might alter Payment.refund_to (if using HTTP, it must be TLS-protected).

This says "must", but what should a client do here if the payment URL is not HTTPS? I suggest weakening this to "should", as sometimes TLS is redundant (e.g. if you're sending to a Tor hidden service).

done.
 
The PaymentACK message contains a copy of Payment, but the BIP doesn't say what to do with it. I assume this means a client is free to ignore it and rely on TCP state to figure out the payment/ack connection instead? It may be worth noting that explicitly.

Added:

"payment | Copy of the Payment message that triggered this PaymentACK. Clients may ignore this if they implement another way of associating Payments with PaymentACKs."
 

In the certificates section, you could observe that "validation" means "verification that it correctly chains to a trusted root authority, where trusted roots may be obtained from the operating system. If there is no operating system, the Mozilla root store is recommended".

Modified that section to say:

"...followed by additional certificates, with each subsequent certificate being the one used to certify the previous one, up to a trusted root authority. The recipient must verify the certificate chain according to [RFC5280] and reject the PaymentRequest if any validation failure occurs.

Trusted root certificates may be obtained from the operating system; if validation is done on a device without an operating system, the Mozilla root store is recommended."


--
--
Gavin Andresen