[comp.mail.misc] Delimiters

stewarte@sco.COM (the termite of temptation) (11/01/90)

Broadcasting live from Mars, here's ace reporter les@chinet.chi.il.us (Leslie Mikesell):

>Picking an "unusual" character combination for a delimiter is sort of
>like saying that the faster cars go through an intersection, the less
>likely it is that there will be a collision.

I don't see the analogy.  I would contend that it's more like saying
that if an intersection has a sufficiently small amount of cross-traffic,
the odds of a collision are low enough not to warrant a traffic signal.
Dunno about you, but lots of intersections where I live are uncontrolled...

-- Stewart
-- 
"I had no message and the message was,
 'we're all Jesus, Buddha and the Wizard of Oz'!"
			-- Andy Partridge (XTC)
/*  uunet!sco!stewarte  -or-  stewarte@sco.COM  -or-  Stewart Evans  */

les@chinet.chi.il.us (Leslie Mikesell) (11/02/90)

In article <11470@scorn.sco.COM> stewarte (the termite of temptation) writes:

>>Picking an "unusual" character combination for a delimiter is sort of
>>like saying that the faster cars go through an intersection, the less
>>likely it is that there will be a collision.

>I don't see the analogy.  I would contend that it's more like saying
>that if an intersection has a sufficiently small amount of cross-traffic,
>the odds of a collision are low enough not to warrant a traffic signal.

Given a fixed amount of traffic through an intersection, if the cars
move faster, the intersection will be open more of the time, ergo the
odds of a collision are lower.  The point was that even though this
may be true, it's not really the way you want to do things.

In the case of mail messages, there are at least 2 "correct" ways that
don't muck with the content.  Store one message per file, or use a
Content-Length: header to encapsulate the body.  The latter method has
the advantage of allowing multiple encapsulations per message as well
as per file by using a Content-Type: Multipart followed by a Content-Length
that includes all the pieces which in turn have their own Content-Length:
headers.

I think SysVr4 is going to use this method, so we might as well start
preparing to see it.

Les Mikesell
  les@chinet.chi.il.us