...His dog translator worked perfectly - all the dogs were running around saying "Hey! Hey! Hey! Hey!"

So this dog goes into a Post Office (Well, we have to assume he finds one of the few left) and asks to send a telegram.

"What you you want to say?" asks the clerk.

"Woof, Woof, Woof, Woof, Woof, Wooooooof, Woof, " says the dog

"Do you realise you could send five more words for the same money?" asks the clerk.

"Yes," replies the dog, but it'd make bugger all sense."

Coat going on even as you groan

If mere people can only judge the canine emotional state correctly 40% of the time then who - or what - was able to determine that the software got it right 43% of the time?

There's clearly a Dr Who angle in there somewhere...

I'd assume that the dog's bark was recorded, and then played back without the video.

You were supposed to tell whether the dog was pining for the fjords (on its own), or really happy (fighting).

Not perfect (the dog on its own could be happy that they've got some peace and quiet finally), but good enough for this sort of science*.

How can these people be called scientists?

As mentioned by David Greaves they could not quantify their results effectively, and the stats are questionable.

I'd like to see clear quantifiable evidence, where humans get it right 100% of the time (at this point the porcine flying league should be well into its 3rd year) as a benchmark.

Is there any word on whether the english language version will include "My hovercraft is full of eels"?

... The Hungarians were barking.

Coat on, door open.

Sigh. I can't resist this.

They're barking mad.

They're barking up the wrong tree.

I'm getting my coat on now.

My experience of dogs isn't great, but 99% of the time they seem to be saying "can I have some food?"

(PH icon because she likes to eat meat too, or so I've been told...)

See also the famous Far Side cartoon "What we say to our dogs .... What they think we're saying". (What they think we're saying turns out to be, more or less, "blah blah blah blah ROVER blah blah blah blah ROVER blah blah blah blah ROVER blah"!)

We were party to a bow-lingual demonstration at the Ig Nobel event (www.improbable.com) in London, part of national science week, a few years ago. A suitable canine participant was brought in and persuaded to take part in the demonstration. The device reported 'I want to help' as the translated bark.

Among other things discussed at the event were farting fish, necrophilia in mallard ducks, and the effect of country music on suicide rates. (Plug: free tickets for the 2008 events are available shortly)

If dogs could talk, the first thing they'd say is "You know, vomit isn't half bad"

It should be doggedly pursued.

I wonder if it would properly translate the barks that my dog emitted while he attempted to cool down his food. Believe it or not, he thought they worked! We (well most of us) know that not barking would have produced the same result.

Patience is a virtue!

Go to http://healingmagichands.wordpress.com/2007/12/18/some-favorite-cartoons/1019/ to see the original cartoon.

IgNobel Prize.

Incidentally, if any intelligent dog learnt to talk he would say "Woof", "Bark", or "Whine" (following Gaspode) to avoid humans from finding out.

- from Dr Evil, Austin Powers.

Anyway, why would I need a dog translator?

If somebody knocks at the door, dogs always make the same barking noise;

'Boo woo woo? Woo woo woof!'

If they are begging for something, it is easy to tell because of the way they jump up onto their hind legs and whine;

'Hgggrrrmmmm? Hmmm?'

If they want to play, they make as if to run away or they attack your feet and bark at you and wag their tails;

'Whaf!'

See?

I forgot to mention that the sound of the door barking woof can be a bit different according to size.

A big dog would woof in the way I described, whereas a smaller dog might bark;

"Waur waur waurrgh? Waur waur waurgh!"

Means the same thing though.

I can't find any articles about this that don't source the Daily Mail, however, by the sounds of it the researchers recorded the responses of the dogs in different situations, then created a programme that, 43% of the time, would correctly give the answer, eg, "the dog that barked just then, he'd just seen a ball", and the human could guess what situation the dog was in 40% of the time.

So, they could easily verify the results the software was giving them, either the dog had just seen a ball, or it hadn't.

Scientific? yes.

Worthwhile? Hmmmm.