spf-discuss
[Top] [All Lists]

Look at that patent application again

2004-09-23 13:38:04
I know I raised this point twice, and was roundly ignored.  Perhaps it is 
because I really am wrong, or perhaps nobody read or understood what I was 
saying.  I am going to say it again.  If you want me to shut up. Please say so. 
 At least I will know then that what I said wasn't just overlooked.

Regarding US Patent Application  20040181571, the one we saw first.

I am not convinced it actually describes or claims the methods used in SPF, or 
even in PRA.  It is very close, but perhaps not close enough.

Why?

Look at where the claims and summary get the network address (IP address in the 
Intenet case) and to the language used to describe that address.

Look at where the claims and summary get the purported sending domain and the 
language used there.

Pay careful attention to the words "actual" , "purported" and "purportedly".  
Look at the phrases "network address" and "actual sending side network 
address".  Also look at the phrase "message parameters.

The way I understand it, SPF Classic and the lamented, but possibly not late, 
Sender ID both take the actual sending side network address from the TCP/IP 
stack, not from "message parameters" as described here.  If so, then it is not 
at all clear to me that claim 22 or its dependent claims or the description 
paragraphs that flesh them out actually describe this situation.  They have the 
algorithm looking at "a plurality of message parameters" in order to determine 
an "actual sending side network address".  I understand that neither SPF 
Classic nor Sender ID has any doubt about the actual sending side network 
address and that neither has any mechanism for finding one in message 
parameters at all, certainly not parsing headers to find it, since headers are 
all untrustworthy.

SPF Classic and Sender ID both parse one or potentially more headers in order 
to find the purported sending domain.  I don't see that action in the 
application.  I do see, in the application, parsing headers to find an "actual 
network address.

Just to help nail down terms.  I take "actual sending side network address" or 
"actual network addres" to mean (in the Internet case) the IP address of the 
sending machine.  I believe the application uses exactly that definition.  I 
take "purported sending domain" to mean the domain portion of an email address 
contained in the message or it's headers (including SMTP headers as well as 
Data headers).  I believe the application uses exactly that definition

SPF Classic looks for a purported sending domain in one SMTP header.  Sender ID 
looks in several Data headers and possibly in one or more SMTP header.
When I read this application, I see it looking in message "parameters" which 
include Data headers, and probably SMTP headers, for an "actual sending side 
network address", **but not for a purported sending domain**.  When it is 
looking at a plurality of message parameters, it is looking for an "actual 
sending side network address".

To me, the patent application and SPF/Sender ID are doing different things.  
Quite possibly different enough that this application doesn't really cover 
either.

Perhaps I am reading this wrong.  I am, after all, using a Microsoft product to 
read this patent application and it is very likely that the omniscient and all 
powerful Mr. Gates programmed this product to obfuscate the meaning of this 
patent application.

Mark Holm


<Prev in Thread] Current Thread [Next in Thread>