Future Of Pedestrian Navigation

Augmented reality navigation (of course).



When can I get the final version of this one for my iphone?

Augmented Reality Comes to the iPhone

The ActiveVision lab has produced some really cool prototype Augmented Reality code for the iPhone:



Flowers growing from books, chasing darth vader around. Almost as cool as the Augmented Reality "Virtual Box Simulator" from the USPS -- it's an AR app that actually has utility in day to day life.

Science Experiment #2 -- Robots Gone Wild

Ok folks, here's my revised hypothesis: All that bit.ly traffic from twitter is just a bunch of robots from real time search engines that are getting a lot of VC energy these days (Ron Conway is investing in 50 of them). So, the experiment is this: I'm going to run my own lightweight URL shortener and look at the Apache logs. Bring on the robots!

Update!

Within seconds of posting this link on twitter, here are some of the bots that came through my URL shortener (sorry for the verbosity):

38.99.68.204 - - [08/Jun/2009:19:34:45 -0700] "HEAD /short/rosenblum/bot/ HTTP/1.1" 302 - "-" "Mozilla/5.0 (compatible; FriendFeedBot/0.1; +Http://friendfeed.com/about/bot)"
216.24.142.41 - - [08/Jun/2009:19:34:48 -0700] "HEAD /short/rosenblum/bot/ HTTP/1.0" 302 - "-" "Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.1) Gecko/20061010 Firefox/2.0 Me.dium/1.0 (http://me.dium.com)"
216.24.142.41 - - [08/Jun/2009:19:34:49 -0700] "HEAD /short/rosenblum/bot/ HTTP/1.0" 302 - "-" "Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.1) Gecko/20061010 Firefox/2.0 Me.dium/1.0 (http://me.dium.com)"
79.99.6.106 - - [08/Jun/2009:19:34:50 -0700] "HEAD /short/rosenblum/bot/ HTTP/1.1" 302 - "-" "Twingly Recon"
66.249.68.121 - - [08/Jun/2009:19:34:50 -0700] "GET /short/rosenblum/bot/ HTTP/1.1" 302 - "-" "Mozilla/5.0 (compatible; Googlebot/2.1; +http://www.google.com/bot.html)"
75.101.138.248 - - [08/Jun/2009:19:34:51 -0700] "HEAD /short/rosenblum/bot/ HTTP/1.1" 302 - "-" "uberbot 1.0"
174.129.89.199 - - [08/Jun/2009:19:34:56 -0700] "GET /short/rosenblum/bot/ HTTP/1.1" 302 - "-" "Python-urllib/2.5"
216.24.142.41 - - [08/Jun/2009:19:34:57 -0700] "HEAD /short/rosenblum/bot/ HTTP/1.0" 302 - "-" "Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.1) Gecko/20061010 Firefox/2.0 Me.dium/1.0 (http://me.dium.com)"
89.151.84.34 - - [08/Jun/2009:19:35:05 -0700] "GET /short/rosenblum/bot/ HTTP/1.1" 302 - "-" "Mozilla/5.0 (Windows; U; Windows NT 5.1; en-GB; rv:1.9.0.6) Gecko/2009011913 Firefox/3.0.6 TweetmemeBot"
216.24.142.41 - - [08/Jun/2009:19:35:22 -0700] "HEAD /short/rosenblum/bot/ HTTP/1.0" 302 - "-" "Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.1) Gecko/20061010 Firefox/2.0 Me.dium/1.0 (http://me.dium.com)"
91.121.89.155 - - [08/Jun/2009:19:37:43 -0700] "HEAD /short/rosenblum/bot/ HTTP/1.1" 302 - "-" "PycURL/7.19.3"
74.125.44.136 - - [08/Jun/2009:19:37:49 -0700] "GET /gogo/atom.xml HTTP/1.0" 301 244 "-" "FeedBurner/1.0 (http://www.FeedBurner.com)"

Science Experiment: Early Results Show Bit.ly Over-Reports traffic By At Least 560%

Well, the results are in (please forgive my possibly statistically insignificant traffic numbers -- my blog just ain't all that popular). Here's a snap shot of what Google thinks happened, and what bit.ly thinks happened over the two day period (May 31, June 1) after my original post and bit.ly tweet:

Google: 24 Total Visits (including 12 from Google Search) to my blog.
Bit.ly: 28 Clicks on the bit.ly link.

In the Google results, there are 5 hits from twitter, which could be the bit.ly link (although that would be a radical interpretation by my understanding, but it is where I get the 560% number), although more likely those are clicks on the link to my blog from my twitter profile. Basically, the bit.ly clicks don't show up at all -- bit.ly is not a referer to my site according to GA.

Here's the Google Result:



Here's the bit.ly metrics:



There are a lot of reasons that Google might not report Bit.ly traffic. A lot of it is perceived as Direct to URL traffic (roughly 80%) -- email and IM clients and the like:



But that should still leave 4-5 visits from bit.ly in my Google Analytics (plus a lot of direct traffic, which is no where to be found). More on this topic as my thinking around it evolves. I assume everyone is seeing this. Anyone have more impressive numbers to report?

Bit.ly v. Google Analytics - An Experiment, Part I

Is there a discrepancy between reporting from different analytics providers? Absolutely. Even comparing log analyzers like AWstats or Webalizer there are major differences in the reporting -- and they are looking at identical source data (log files).

It gets tricker with the introduction of URL shorteners like bit.ly and tinyurl.com, as they are adding a layer of indirection, proxying all traffic through a single referer. Add the complexity of tracking referer traffic from twitter (where a large percentage of traffic is from API usage, and therefore might appear as 'direct to url' traffic in Google Analytics), and it's awfully hard to find the truth in analytics.

On top of that, @wiseleo notes: "A HEAD request instead of a GET request, which is a common scenario when expanding compressed URLs, can create ... discrepancies." (although bit.ly in particular doesn't count HEAD requests in their analytics).

Bob noted that he saw a lot more traffic in his bit.ly results than in Google Analytics for his widely tweeted article on how Twitter Loves You More Than Oprah. So, here's an experiment. I'll try to get as much traffic to this article via bit.ly as possible, and then run my own comparisons. Let's see if we can hone in on the 'bit.ly effect' on analytics data. Generally speaking, my answer to this conundrum has been to choose a source you trust (i.e. WebSideStory, Omniture, Google Analytics, or what have you) and use that as your canonical reference. But it's always good to test your assumptions, and recheck regularly.

I'll post a follow up when I have some data :)

[update: see results of the bit.ly and google analytics comparison]

Mixed Reality Gaming: Working My LevelHead Cube

Check out LevelHead (awesome open source mixed/augmented reality game/art project) in action:

Augmented Reality USS Enterprise

I don't have a PC with a webcam for the enterprise experience, but you can get a good sense for the experience in this video.