Friday, October 24, 2008

Please return item to the bagging area, revisited

Dan broke the bad news last week that these satanic machines have spread around the world.

If you recall, my voyage into hell began when I tried to use a supermarket self-checkout machine and couldn't get the thing to stop shouting at me.

After my most recent battle, I decided to do something very unusual: make eye contact with and speak to another person in a busy shop in London.

No. I hadn't gone out of my mind... this person worked there.  But still.

I asked why the machine was so demonstrably rubbish and he made a useful observation: the machine takes some time to register that an item has been placed in the bagging area. 

This means that if you want to remove a bag, you have to wait for the item to be registered and then remove the bag.

I tried his idea and, sure enough, if I waited a second or so, the item would be registered. Then, if I lifted off the bag, it would still shout at me but one of the options was to tell it that I had lifted off the bag.

Phew!

I think I also understand why the machines work in this way.  The question is... is there a better way?

Consider the design meeting.

Requirement: "All our customers are thieves so we need to be able to detect if somebody places an item in the bagging area if they haven't scanned."

Now... an "obvious" solution to this is to put in a sensor that reacts when an item is placed in the bagging area.  If an item is registered without a preceding scan operation then something dodgy is going on.

So far, so good.

But there's a problem: different customers will scan in different ways.

Sure... some of them will "scan, bag, scan, bag, ..." and this detection mechanism will work.

But some of them will pick up three items, scan all three and then put two of them in the bagging area and hold on to the other one because they know it's fragile and they want to put it in last... and then they'll scan two items and put both in together at the same time.... (you get the idea).

This means that it would be impossible to tell whether somebody had placed an unscanned item into one of these batches and the requirement could not be met.

Now, the solution they chose to employ was simple and cheap. They simply enforced the "scan, bag, scan, bag, ..." approach in the most clunky way possible.

However, it's interesting to think about what else they could have done.

One user-friendly idea could have been to add a set of scales to the side of the machine where the unscanned items are kept prior to scanning.

If the total weight of the "to be scanned" items equals the total weight of the "scanned" items at the end of the process then they would know there had been no skulduggery.  (I can immediately think of a few problems with this approach, however...)

Either way, I still hate those machines.

Friday, October 10, 2008

My faith in the Economist is restored!

I wrote last week that the Economist had cancelled my subscription without warning.

Well, they're back in my good books now. Their PR people saw my posting and somebody from The Economist has just been in touch to sort it out.

An economic illiteracy crisis in East London has just been averted!


Sent from my BlackBerry® wireless device

Friday, October 03, 2008

The Economist May have lost a subscriber

I switched bank a few months ago. The new bank (First Direct) contacted all the companies that I have a Direct Debit with and told them to move the mandate from my old bank to the new one.

They all did it.

Apart from one:

The Economist.

They seem to have cancelled the old Direct Debit but not bothered to set up the new one.

Worse, rather than contacting me immediately when the last payment failed to go through, they cancelled my subscription and sent me a letter instead.

Which means this week's copy hasn't been delivered.

:-(