Floyd L. Davidson wrote:

>   2) Prefix each line of quoted text with a '>' character, and
>      mark each level of quoted text with an initial attribution
>      line that identifies who authored each level.  That results
>      in a long running exchange that follows this form:
>
>          John Doe III wrote:
>          >Jane Doe III wrote:
>          >>Joe Blow wrote
>          >>>Jane Que wrote
>          >>>
>          >>>>My name is Jane,
>          >>>>the sun is up.
>          >>>>
>          >>>Well my name is Joe, and
>          >>>the sun is down here.
>          >>>
>          >>My name is also Jane,
>          >>what is a sun?
>          >>
>          >I'm John Doe, who
>          >is Jane?
>
>          My response to all of that goes here.
>
>    3) Not obvious from item 2) is that this format is applied at
>       a minimum to _paragraphs_, but often to sentences or even
>       just to individual lines or fragments.  Very rarely should
>       there be multiple paragraphs retained as quoted text with
>       a comment at the bottom referencing anything more than a
>       paragraph above where it is placed.

I always wondered if an xml-like notation might catch on for this
instead. Yes, it's nice when your email client automatically adds the
'>'. And it look pretty good. But it certainly bites when that doesn't
happen or the lines get broken up in a odd manner. I would think
something like:

  <quote author="Floyd L. Davidson">
     2) Prefix each line of quoted text with a '>' character, and
        mark each level of quoted text with an initial attribution
        line that identifies who authored each level.  That results
        in a long running exchange that follows this form:
  </quote>

Would be much easier for email clients to work with. They could easily
"pretty-print" these sections. 

T.