By: Steven Citron-Pousty (a)
Created: 1/2/08

This "press release" was just darn fluffy not to have some fun with.

I mean the title alone is so over the top - Appistry Demonstrates World’s Fastest Geocoding Solution - did they call in Guiness Book on that one?

 

Of course this made the rounds, since who doesn't need a good laugh in the New Year. Given the lack of details in the release about things like say accuracy of results, structured versus unstructured geocoding, if they are all valid addresses, the speed or memory of the machines...  We already had a couple engineering solutions that beat the pants off the world's fastest geocoding solution:

 

Run the benchmark against a dataset with all the addresses in the U.S. pre-geocoded

 

and my favorite (in one line of perl no less)

 

while(<stdin>) { print "0.0 0.0\n"; }

 

 

 How you like me now! You ready for the press release on this one...

So if they really want to throw down the gauntlet on speed of geocoding why not actually publish some real specs and details on what they did. I am sure Appistry make some nice cloud computing software but when talking about benchmarking we could use some more details...