Re: [Exim] MX Record points to non-existent host

Pàgina inicial
Delete this message
Reply to this message
Autor: Exim Users Mailing List
Data:  
A: Greg Louis
CC: Exim Users Mailing List
Assumpte: Re: [Exim] MX Record points to non-existent host
[ On Saturday, February 22, 2003 at 11:31:30 (-0500), Greg Louis wrote: ]
> Subject: Re: [Exim] MX Record points to non-existent host
>
> It is not my duty -- I would say it is none of my business -- to get in
> anyone else's face about how well or badly they set up their
> environment.


Well, actually it is. They don't call this thing an "INTERNET" for
nothing!

For every problem you tolerate you create at least two more problems for
others, some of whom are simply not able to tolerate the same kinds of
problems you think you can tolerate. At least one of those new problems
is far more critical because it is not a technical problem.

Interoperability standards exist for very good reasons.

The interoperability standards we have now are widly believed to already
be the absolute bare minimum necessary. I.e. no more tolerance at a
general level is possible (or at least more tolerance would be extremely
unwise). One can certainly easily imagine far more stringent standards
(just look at international telecommunications!). The reasons for every
rule are not always so obvious, but there are almost always situations
where some people simply cannot "work around" some things, at least not
without throwing out everything they have and starting over with new
software -- meanwhile everything they have will continue to interoperate
100% with every other fully conforming software.

By tolerating some deviation from the interoperability standards you
create an impression and even an expectation that the standards are
irrelevant and that the tolerated behaviour is actually OK. Many of the
rest of us in the Internet community do not appreciate this at all.

> If the users of such lists choose to deprive themselves of the ability
> to communicate with such hosts as do not meet their standards, that is
> their decision and their right. Our need is the exact opposite: to
> communicate with anyone who wishes to contact us.


Do you expect that your phone will ring whenever someone wants to call
you no matter what number they dial? Or vice versa?

Do you expect your FAX machine to produce a facsimile of the document
fed to it if the FAX machine at the other end has had some new
non-standard high-power compression algorithm installed in its firmware?
Or vice versa?

The only difference between our Internet interoperability standards and
the more formal standards that make the telephone system work, that make
FAX machines and modems from every manufacturer talk to each other,
etc., is that our standards are made just as informally as our network.
However our standards are no less important. If you operate a strictly
RFC-compiant infrastructure then you are already allowing anyone to
contact you if they wish.

Using your ability and willingness to ignore interoperability standards
as a sales tool to attract business from those who are ignorant of those
same standards really angers the rest of us in the Internet community,
even, and maybe especially, those of us who are not directly competing
against you for the same business. If you would instead use your skills
and knowledge to help those same people then we would applaud you
instead (and whether your generosity would give you a competitive
advantage in your business depends only how you go about deploying your
skills to help your customers).

> You live in a different business world from the one I inhabit. We do
> not criticize our customers' networking skills any more than we would
> their body odour. I do sometimes address network issues with business
> partners, if I know they have the ability to understand and the
> motivation to examine my submission -- not with customers.


This isn't about criticism -- this is about very basic minimum
requirements like making sure everyone's computer is actually plugged
in, not just turned on; about making sure the buttons generate the right
tones on everyone's telephone; about making sure all the FAX machines
adhere to the international standards for FAX communications, etc.

You can jump up and then ask how high as you're in flight for every
demand your customers put on you, but at some point if several of them
are continuously telling you different heights at the same time then
eventually you're going to be torn to bits. You cannot survive being
tolerant all the time of everyone's idiocy, laziness, and bull-headed
ignorance. If one of them doesn't eventually get you then one of us
will! ;-)

If you tell your customers that your FAX machine is a standards
compliant FAX machine that is guaranteed by its manufacturer to talk to
every other standards compliant FAX machine then I'm sure they'll stop
trying to send you FAXes using their new proprietary hyper-compression
FAX machine no matter how much they might like it. If they offer to
sell you a version of their FAX machine then you can politely decline
and tell them you can't afford it. If they offer to put one in your
office free of charge then you'll have to ask your CFO if you can afford
to operate an additional custom machine based on the profit that will be
generated by continuing to be able to do business with that customer.

The same applies for your mailer. If you tell your customers that your
mailer and your DNS infrastructure meets not just the letter of the law
in the RFCs, but also their intent, and that you can send and receive
e-mail between every other RFC-compliant mailer then I'm sure they'll
realise that they're the ones who are causing the interoperability
problems. However if you offer, either directly or indirectly, to spend
a bit of effort to allow your customer to continue to operate their
software in a non-compliant manner then you'd better watch out because
if someone like me finds out that you've been giving false expectations
to other Internet community members then I'm going to make absolutely
certain that you're the one who gets tarred and feathered for their
problems. Your customers are going to be a lot more angry with you in
the long term than if you had just given them a bit of good advice in
the first place.

(Note there's a universe of difference between claiming you can send and
receive e-mail with many other domains and claiming you can send and
receive e-mail with RFC-compliant domains. Never ever make the mistake
of claiming you can interoperate without knowing exactly what you can
interoperate with. I'd like to do very nasty things to every idiot who
has ever claimed they could send and receive e-mail with every other
domain but mine when the problem is clearly their own! Locking them in
a bare 1800's jail cell with nothing but printed copies of the RFCs
would only be a good start.)

--
                                Greg A. Woods


+1 416 218-0098;            <g.a.woods@???>;           <woods@???>
Planix, Inc. <woods@???>; VE3TCP; Secrets of the Weird <woods@???>