Larry Jaques wrote:
Who IS pulling all those global strings, anyway? You's a
gummint man. Fess up.

There's this guy.

OK, I've told too much.

UA100

In article ogers.com,
says...
On 28-Oct-2003, "Edwin Pawlowski" wrote:

Perhaps they confirmed "receipt" of the order, but not necessarily that it
would be shipped or that your CC has been approved.

Maybe the web page programmers botched it and issue a confirmation
by mistake. After all, how much software is actually reliable?

Getting one's shorts in a knot over this issue seems to me a
waste of energy. I've missed lots of opportunities - I get
over it.

Mike


Actually, I'm fairly certain that what they did was to set up their
system to remove things from inventory when the purchase was completed,
and not when the customer placed it in their cart. It's all water under
the bridge in any case. I have a local Rockler's that is more than
happy to take my money.

On 30-Oct-2003, kenR wrote:

Actually, I'm fairly certain that what they did was to set up their
system to remove things from inventory when the purchase was completed,
and not when the customer placed it in their cart.

If they did that and the customer never completes the transaction, then
the item stays out of inventory even though not sold. They'd have to
design a timeout mechanism to put it back. Too long and they miss
inventory - too short and the customer finds his order suddenly
disappeared when he went to the bathroom. The web is stateless
and that makes foolproof transaction processing difficult at best.

Mike

I don't buy that. I cannot think of a single modern relational database that
does not have transactional processing built in ... this should not be a
problem/excuse, even in a stateless environment like the web.

It's done all the time. If what you say was the case, there would be no
Internet banking over the web.

--
www.e-woodshop.net
Last update: 9/21/03


"Michael Daly" wrote in message

If they did that and the customer never completes the transaction, then
the item stays out of inventory even though not sold. They'd have to
design a timeout mechanism to put it back. Too long and they miss
inventory - too short and the customer finds his order suddenly
disappeared when he went to the bathroom. The web is stateless
and that makes foolproof transaction processing difficult at best.

Mike

Sorry, Mike ... I didn't mean that to sound so snippy. Had to work late and
still on the first cup of coffee. Mea culpa.

--
www.e-woodshop.net
Last update: 9/21/03


"Swingman" wrote in message
I don't buy that. I cannot think of a single modern relational database
that
does not have transactional processing built in ... this should not be a
problem/excuse, even in a stateless environment like the web.

It's done all the time. If what you say was the case, there would be no
Internet banking over the web.

--
www.e-woodshop.net
Last update: 9/21/03


"Michael Daly" wrote in message

If they did that and the customer never completes the transaction, then
the item stays out of inventory even though not sold. They'd have to
design a timeout mechanism to put it back. Too long and they miss
inventory - too short and the customer finds his order suddenly
disappeared when he went to the bathroom. The web is stateless
and that makes foolproof transaction processing difficult at best.

Mike

On 31-Oct-2003, "Swingman" wrote:

I cannot think of a single modern relational database that
does not have transactional processing built in

The relational database is hooked to the web server, not the
user. How does the relational database know what the user is
doing if the transaction takes several _independent_ steps?
Completion/rollback of a transaction has to be triggered from
something - what is that?

Mike

The mechanism varies. It could be as simple as a cookie, or a timeout, or as
complicated as a client side script and buffers. "Commit trans" is never
sent to the database without a specific, verifiable trigger mechanism ...
as I stated earlier, Internet banking would not be possible unless there was
a mechanism to do this in a stateless environment.

Hell, commerce on the Internet would be impossible if there was no
mechanisms for maintaining state on a web transaction.

--
www.e-woodshop.net
Last update: 9/21/03


"Michael Daly" wrote in message
On 31-Oct-2003, "Swingman" wrote:

I cannot think of a single modern relational database that
does not have transactional processing built in

The relational database is hooked to the web server, not the
user. How does the relational database know what the user is
doing if the transaction takes several _independent_ steps?
Completion/rollback of a transaction has to be triggered from
something - what is that?

Mike

"Larry Jaques" wrote in message
...
Oh, that's right. I forgot that it was -they- (our CIA) who
created Khadafy (Syria), Hussein (Iraq), the Shah (Iran),
and most of the other terrorists who are around today.

Khadafy is in Syria now? Man, I new some of these guys were nomadic, but
this is news - who's running Libya then? No, wait, who's on first isn't
he....

Greg