clipka <ano### [at] anonymousorg> wrote:
> Am 21.06.2012 16:58, schrieb Invisible:
> >>> Not really - the search keyword "Vidocq" /would/ have helped...
> >>
> >>> <ducks& runs>
> >>
> >> That's not a search keyword. That's the answer.
> >
> > It's like an NP decision problem; the solution can be easily verified,
> > but not easily computed.
> >
> > (Except that it's not a decision problem. And we've not talking about
> > computation, we're talking about Google searches. And that this is a
> > horrible analogy...)

> Maybe deep down at the core of the problem it /is/ an NP decision problem?

  Not really. If the input is all the movie covers in existence, then a
linear search would be enough to find the described one.

-- 
                                                          - Warp

Post a reply to this message

Am 21.06.2012 18:05, schrieb Warp:
> clipka<ano### [at] anonymousorg>  wrote:
>> Am 21.06.2012 16:58, schrieb Invisible:
>>>>> Not really - the search keyword "Vidocq" /would/ have helped...
>>>>
>>>>> <ducks&  runs>
>>>>
>>>> That's not a search keyword. That's the answer.
>>>
>>> It's like an NP decision problem; the solution can be easily verified,
>>> but not easily computed.
>>>
>>> (Except that it's not a decision problem. And we've not talking about
>>> computation, we're talking about Google searches. And that this is a
>>> horrible analogy...)
>
>> Maybe deep down at the core of the problem it /is/ an NP decision problem?
>
>    Not really. If the input is all the movie covers in existence, then a
> linear search would be enough to find the described one.

I guess it's more complex than that: You don't want /the/ match (because 
your question is too fuzzy for an exact one), but the /best/ one.

The metric for the quality of the match would have to include contextual 
analysis; and this analysis needs to be more fine-grained the more 
candidates you have.

Post a reply to this message