PE> BL>   Have you actually written your reader yet? How do you do it? You
PE> BL> just put EOT ahead of the Tearline... right? So what's EOT doing? If

PE> Bob doesn't even know what a tearline is yet, so you can safely
PE> ignore anything he says on the topic until then.

FM> I think I know what a tearline is, but I don't see why it enters this
FM> discussion. It contains no useful information does it?

It contains the same sort of information as is on the PID.  Whether
you think it is useful to know what software created the message is
up to you.  I just want to recognize it and not display it, for the
same reason I don't display the PID.  Or the SEENBY lines.

PE> Yes, if tearline is compulsory, AND a blank line before the tearline
PE> is either mandatory or forbidden (by software automatic-generation),
PE> then yes, there is no need for EOT.  As I said a long, long time ago.

FM> But, OTOH (and this is my point), if you are the originator of a message
FM> you can put that other stuff in, where it belongs, all the time. So for
FM> a message originator there's no need for EOT, and for a subsequent
FM> reader you can't reliably add it. Mostly you will be able to *correctly*
FM> add it, in those cases where you don't need to.

"need" = "able to distinguish which bit is user-text and which bit
isn't".

PE> BL> If found look for an EOT. If found remove it and add your own in front
PE> BL> of the Tearline. The other dopey bastard may have put his EOT in the
PE> BL> wrong place.

PE> Yeah, right.  Can't rely on them getting EOT right, so best not
PE> rely on them getting the origin right either, so ignore that.  And
PE> ignore the text too, and the header, yeah right, makes perfect
PE> sense now.

FM> As someone who's trying to learn and understand this stuff, I'd like to
FM> see the reasoned comments on both sides.

I'm not sure if you're implying that my above comment is not
reasoned.  It is reasoned, but maybe it was too subtle:

1. Bob is saying that you can't rely on EOT because the EOT-generating
software may not be following the spec.

2. That is completely illogical, because you can say the same thing
about any of the control lines, INTL etc etc, don't use them in case
they're out of spec.  Software is not about not using control 
information on the offchance that it is out of spec, you are meant to
assume that it is in-spec, and it's the sender's own problem if they
stuffed up and sent their netmail to Africa instead of America, or
said that the user had written a line starting "---" when they hadn't.

3. The above two points are self-evident, I don't know how to explain
in lower detail than that.

BFN.  Paul.
@EOT:

---