Re: Content-Length harmful

[Michael Stone <mstone@itri.loyola.edu>]

Quoting Andrew Pimlott (andrew@pimlott.ne.mediaone.net):
> > Since this is Debian, we could add policy that ALL Debian MUAs must honour
> > the Content-Length (maybe they all do anyway) and MTA's should be set up to
> > expect this.
> > 
> > I'm not sure if it's a good idea or not.
> 
> Some MUA authors don't think so.  Something about being liberal in what you
> accept.  See
> 
> http://www.washington.edu/pine/faq/sysadmins.html#xtocid1859727

I'm inclined to discount anything coming from uwash in the MUA
department (I loathe pine.) Their major argument seems to be that it
will mess up the performance of their code (Pine? performance?), though
they do say that "There are many serious technical problems with the
Content-Length: header" without specifying much about what they are.
They also suggest that you don't quote every "\nFrom", but only certain
ones (as if that's 100% guaranteed on a system where you don't control
the software; if we can dictate a particular From format, we can also
dictate Content-Length compliance.) The only argument I see that holds
water is the vi-editing one, to which the only reply can be "don't do
that"...we don't recommend people trim executables with vi, either.

> http://www.netscape.com/eng/mozilla/2.0/relnotes/demo/content-length.html
> 
> The latter is a classic.

I think that the best comment was at the end: "You can't win." I prefer
Content-Length for the simple reason that it's the easiest way to get
unmunged mail in a near-mbox format. If anyone has a better solution,
I'd love to hear it...

Mike Stone