[ietf-dkim] on the suitability of the From header field

[Keith Moore]

domainkeys-feedbackbase02 at yahoo.com wrote:
>>If end-users today are accustomed to thinking the message was sent by 
>>RFC2822.From, they will need to be educated, and they may also need 
>>better MUAs that make the distinction clear.  But I don't think most 
>>end-users are this naive or incapable of understanding the difference. 
> 
> 
> Sure. But do end-users want to, and should they have to?
> 
> Just because a rickety old standard technically makes it possible to
> distinguish between a confusion of originating, authoring, authorizing,
> resending and non-delivery report addresses, does not imply that it should  be
> imposed on the billion or so end-users who have little choice but to use that
> standard.

And just because you prefer a simple-minded view of communications 
doesn't mean that such a view should be imposed on the billion or so 
end-users who use email.

People using snail mail have little or no trouble understanding the 
difference between the name and address on the outside of the paper 
envelope, the name and addresss at the top of the letter, and the name 
of the person who signs the letter.  Each of these names/addresses 
serves different purposes which are easily understood.  Email is no 
different, and no more difficult to understand.

Existing email user agents might blur the distinction between these, but 
this is a user interface problem rather than a protocol problem.  And 
user interfaces _will_ have to adapt to take advantage of email 
authentication, because email authentication creates new conditions that 
users will need to be able to distinguish from existing conditions.

> It's also an interesting presumption that it is the end-user that needs
> educating about the, er, richness of email addressing.

Go back and reread what I wrote.  I don't actually presume such.

> Perhaps it's more important for us implementors to be educated about how end-users really want to
> think about email addressing and adapt our thinking to their needs?

Fine.  Just don't try to lump all end-users in the same bucket, and 
don't insist that the mail protocol be crippled in order to accommodate 
the most simple-minded of end-users.  Email needs to be able to serve a 
wide variety of users.

> Personally, I'm developing the suspicion that end-users have been using email
> long enough now that we should be noticing the demand for complex
> originating/sender addressing - especially in such a competitive and diverse
> marketplace. But at least from where I sit, the evidence is slim at best.

We're certainly seeing wide and quite legitimate uses of differences 
between header From and return-path. We're also seeing increasing use of 
subaddresses.

Keith