[RFCI-Discuss] conservative sending versus liberal receiving

[Anthony Howe]

James Ralston wrote:

> On 2004-09-29 at 19:33+02 Anthony Howe <achowe at snert.com> wrote:
> 
> 
>>In recent years I've come to believe that the mantra:
>>
>>        Be conservative in what you do,
>>        be liberal in what you accept from others.
>>
>>Has actually hurt the Internet in the long term. Its because of this
>>mantra that programmers have been lazy or not strict about RFC
>>conformance.  As a result viruses and spammers abuse this liberal
>>ideal.
>>
>>To put it another way, if we had been strict about conformance in
>>the early days of the Internet, when there were fewer
>>people/servers/MTA software, we might have less of a mess today,
>>because the discipline would have been instilled early.
> 
> 
> You're correct.  If we had been strict about conformance in the early
> days of the Internet, we wouldn't have a mess today.  Rather, the
> Internet would have shattered into balkanized fragments, withered, and
> died.

I don't think so. In the early days, such problems with RFC conformance 
would have been addressed with revisions, corrections, refactoring, and 
fine tuning.

The problem is that the fine tuning process has fallen way behind the 
curse of main stream growth and behaviour. So only now are we seeing the 
result of a liberal protocols designed for a like minded community. I 
miss those days, but now we have to deal with it. RFCI is one approach, 
which I hope will one day go away when a new SMTP RFC comes about and is 
  100% adopted (Dreamer. Nothing but a Dreamer...).

>...
>     (e.g., not object to technical errors where the meaning is still
>     clear).

This is where I have my dispute. Technical errors are significant and 
its this loop hole that provides wiggle room for spam and viruses. Yes, 
ideally in the beginning you wnat to be liberal, but you can't remain so 
forever, you have to keep improving and refining the process such that 
there should be no technical errors.

Too many people have settled for good is good enough, instead of moving 
towards above average and excellent. They, the developers, that are 
suppose to implement the RFC standards often rely on the liberal 
attitude of the receiving end, because they're too lazy to finish the 
job properly.

> Case in point: 8-bit characters are illegal in RFC2822 message
> headers.  Always have been.  But in the pre-MIME days, if you wanted
> to receive mail from people who couldn't express their alphabet in
> US-ASCII (meaning, the vast majority of the people of the world), you
> accepted 8-bit characters in both the headers and the body.  Period.

> (As an aside, I remember at the time how hard some people fought
> against MIME.  Many Europeans wanted to go right on shipping arbitrary
> 8-bit data around, and objected strenuously to quoted-printable
> encoding.  That MIME succeeded is a testament to how hard a lot of
> people worked to not only draft protocols, but to build bridges.)

I agree, a technical issue was refined and corrected, but now you still 
have people failing not to follow RFC 2045..2049 and still send 8-bit 
characters in headers in light of there being RFCs that stipulate the 
correct format for messages. I even wrote a milter to address that, 
because its becoming so annoying.

> Spammers and viruses will only "break" the Internet if we are foolish
> enough to abandon the "be conservative in what you send and liberal in
> what you receive" philosophy in order to combat them.

I see the opposite in that; the failure of the Internet to move forward 
and implement standards and demand conformance will just allow the 
current situation to get worse, because they continue to abuse the 
liberal nature of Internet protocols. At some point you have to tighten 
the screws else the box will fall apart.

-- 
Anthony C Howe                                 +33 6 11 89 73 78
http://www.snert.com/       ICQ: 7116561         AIM: Sir Wumpus

            "Once...we were here."  - Last of The Mohicans