it seems like there are better ways for lists to tag their contents
and for MUAs to display information about list origin to recipients.
I agree. But I don't think lists are the only problem here.
More generally, the problem is that there are lots of things that think
they should be able to demand the recipients' attention, and there are
only two places they can reliably do so - the subject field and the
message body. And for this reason both of them commonly get altered.
Now let's say (as a thought experiment) that we had the luxury of
changing how MUAs display messages. Then we could define a couple of
new fields:
"Display-Header-Fields:" field-name *("," field-name)
which would go in a message header, and would say to the MUA that the
indicated field-names are worthy of display to the recipient in
addition to whatever other fields were normally displayed.
"Content-Added-By:" identifier
which would go in a body part header, and would indicate that the
content was not supplied by the message author but was instead added by
some intermediary. Content-Disposition fields could then indicate
whether the intermediary thinks the additional body part should be
displayed "inline", or made available as an attachment, or what.
This would make it possible for the recipient's MUA to separate
annotations from original content.
Now the question is - would the intermediaries be willing to mark their
annotations in such a way that the content could easily be separated?
Some lists would, undoubtedly. Maybe virus filters also (though I
think they're more likely to remove content than to add it). But what
about things that add advertisements?
The transparency issue seems to crop up at every layer and in every
popular application. It's as if there is a battle between the
intermediaries and the endpoints. Either the intermediaries think they
are smarter or know better than the endpoints (as in the case of
security firewalls and virus filters), or they think that they have the
right to impose demands on the endpoints that involve changing the
content exchanged by the endpoints (as in the case of "walled gardens"
and "value added services" implemented on top of what were intended as
end-to-end protocols). In every case the result is a arms race where
the intermediaries and endpoints both become more and more complex over
time -- until the system as a whole becomes so expensive and/or
dysfunctional that there is no longer sufficient incentive to fix the
problems.
Keith