At 5:53 AM -0500 1/30/1997, Dr. Manion wrote:
>This wasn't a bash on AOL this was a statement of fact. If AOL was
>truly responsible it wouldn't have taken the threat of court action to
>get them to take action.
I've got this *wonderful* little joke I found at work. When I
get there Monday morning, I'll forward it to myself at home, then
onto the list. However, let me summarize it here:
You see, this guy threatens to sue the telephone companies,
because sometimes he tries to call someone (especially on a holiday
like Mother's Day), and all he gets is busy signals. Obviously, the
phone company is incompetent and should be sued for selling him a
service he can't use.
Likewise, he's going to sue every government on the face of the
planet, because when he goes out onto the highway at rush hour, he
has to wait in traffic.
He's also going to sue every bank and other financial institution
in the world, because they obviously should have increased their
capacity so that he wouldn't have to wait in line.
In addition, he's going to sue the cable companies for all those
times when he turned on the set and the reception wasn't perfect, or
when he had to wait on the phone while they processed his order for
some pay-per-view service.
Suffice it to say that it puts into very real perspective what
it's like for people to file suit against us for not having
sufficient capacity.
However, don't misunderstand me -- I'll be the first person to
speak up and say that AOL screwed up *royally* in not having more
hardware in place sooner. Unfortunately, that was a decision made
somewhere above the level of the people who have to try to manage
this system (i.e., all us technical guys who had some clue as to how
ludicrous this concept was), and we have just had to live with it.
Now, you try working fourteen-plus hours a day trying to hold a
system together with your bare hands, and we'll see how you react
when you see the kind of AOL-bashing I've seen here.
I guarantee you that everyone in Operations is literally working
their butt off (and endangering their marriages, etc...) to keep this
system up and working, and more importantly, to expand the system
quickly enough that we can catch up to and overtake the growth we've
seen, so that we can get ahead of that damn eight-ball.
If you want to help us in that process, please do. If you have a
problem with some part of the system right now, if you bring that to
our attention in a reasonably professional manner, we'll do
everything we can to help you get that problem solved (given the
physical constraints of how many seconds there are in a minute, how
many minutes there are in a day, how many days there are in a year,
how fast our vendors can roll hardware out their doors, etc...).
>There are all kinds of excuses that AOL has for their actions. They
>have the resources to first run "controlled tests" of their email
>before providing it. Whatever problems their email or subscribers
>cause is still their responsibility *not* the responsibility of the
>listowners who are providing service to everyone.
We do controlled tests. In fact, from what I know of the sizes
of the other largest systems on the 'net, our *test* systems are as
large as or larger than the *production* systems of most of our
competitors.
However, that fact means that there are some issues of scale that
simply will not show up until something is rolled out into
production, because there is nothing else on the planet that is of
the same scale and type.
In addition, no matter how big your test systems are, there are
simply some things that are esoteric enough that they cannot possibly
all be tested. If that were the case, then nothing but perfect code
would roll out of Microsoft's offices in Redmond, because they've got
more money to spend on testing their various OSes and applications
than anyone else in the world, and yet their code is buggier than
most anyone else I can think of.
I can only assume that they don't really care, or they could at
least do better than they currently manage (knowing something of
operating and testing very large systems, I would never expect them
to be perfect).
>AOL is a commercial service and despite illusions of grandeur,
>doesn't own the Internet. I personally wasted three days cleaning up
>after AOL. Three days which could have better spent elsewhere.
In all likelihood, if you compare the amount of time spent
cleaning up per user, we're not costing you much more effort or money
per users, it's the fact that we have 20% of the entire Internet
population on AOL. I'd say our "error" rates per capita are probably
about as low as or lower than most any other site on the 'net, it's
just that we have so many more users than any other site on the 'net.
At least, I say this based on my own experience as a list manager
(in my private life), and on my conversations with other list
managers (over the years). If this is not true for your particular
list, then there must be something unique about your list that is
causing you more problems than most others.
--
Brad Knowles, MIME/PGP: brad@his.com
comp.mail.sendmail FAQ Maintainer <http://www.his.com/~brad/>
finger brad@his.com for my PGP Public Keys and Geek Code
The comp.mail.sendmail FAQ is at <http://www.his.com/~brad/sendmail/>
Follow-Ups:
|
|
