What is the smallest number ungoogle number, i.e. number that cannot be found on Google? Hjalmar Gislason came up with this intriguing question. He estimates it is somewhere around 2 million and has a candidate slightly bigger, 9483287. Of course tomorrow, this number will turn up two hits.
How to find this number? Well, a brute force approach is one option, but trying to do 2 million searches at once is going to be noticed at Google and will hardly be within the terms of use of our new overlords and to be kicked by Google and not be able to search again would be bad. Google does 200 million searches a day, so we're talking about 1% of Googles alleged 30 000 servers, i.e. 300 servers running one day.
We probably could estimate the number by first sampling random numbers and put them on a logarithmic scale and then find where a line through these number would hit 0.5. Then search around this point and we might get lucky. But we would never be sure until we tried very single one uptil this number. And then I would publish it and Google would spider it and it would all be in vain. Of course we could then call this number the douweosinga.com number, because it only appears on my blog (until somebody quotes this).
Using a Google key, you can do a 1000 searches a day, so it would take about 6 years to do 2 million searches. Where do these big numbers come from? Most of them are telephone number or part numbers. If you think about it, a lot of telephone numbers are 7 digits and with the enormous amount of area codes in all countries, the chance of a telephone number not existing anywhere is slim. (it is always fun using Google to find out who has the same telephone number as you do).
A distributed approach is probably the only thing that makes sense, but the distributed computation space is already filled with lots of more usefull things like finding aliens or cures for cancer & AIDS. So maybe if somebody at Google is reading this, they could have a look at the index and mail me the number.
0 comments:
Post a Comment