Search Bias, Preferential Placement, User Behavior, and Misleading Statistics

I read this blog post “Does Google Favor its Own Services?” on AlleyInsider and was annoyed by the misleading statistics and interpretation that are written. Here are the reasons this article frustrates me:

  1. Google should be excused for providing their own service in their own top search results if another search engine also provides it as a top result (Bing also returns http://maps.google.com as a top result for example). That would seem to validate that Google Maps is actually a top result, not (just) Google promoting its own service. By my count this invalidates 3 of the 7 dings against Google.
  2. This article should mention (if not account for) that users searching on Google are probably inclined to be more likely to be searching for Google services. It makes sense that a user trying to find the login page for Google email will go to Google search to try to find it. It then makes sense that Google (or its search algorithm) would account for a frequent user behavior (searching for “mail” and then clicking through to mail.google.com) by prioritizing prioritizing the result for mail.google.com in the future (in general, or on a per user basis).

The article should be pointing out that Yahoo seems to be the most out of whack with the other two sites (look at the results for “calendar”), but Yahoo isn’t going to get as many page views as bashing Google in the headline of the post. I hate that AlleyInsider does this all the time, but it got me to read and write about the story so I guess sensationalist blogging wins out again.

My understanding was that Bing “provides” search for Yahoo now anyways, so I’m curious as to what’s going on that Yahoo would still offer different results. Perhaps this is in the line of thinking of trying to accommodate Yahoo users searching for Yahoo services.

Perhaps we should be skeptical of the results Google (or any other search engine) returns, but also keep in mind that algorithms can be written to help create the best experience for Google users (looking for Google services) rather than a vanilla internet user. Now if Google in some way subverts my obvious preference for something else that is more dubious, but perhaps in that case we users should cast our votes for disapproval by not using their service.

Do you “trust Google’s results?” Am I crazy? Am I wrong? I imagine I’m not any more wrong than AlleyInsider in this case.

Panoramic Sweep on the Sony WX1

I recently drafted up this quick little email to some co-workers looking at capabilities of the Panoramic Sweep feature on the Sony WX1.  I bought this little point and shoot to take with me on my trip to South Africa for the World Cup.

Here’s what I learned taking pictures with the PanoSweep feature on the WX1 (a.k.a. the Taylor Swift camera).

It took some practice but after I got the hang of the sweeping motion the shots came out pretty well.  I thought this Korean flag was pretty awesome:

2010-06-22 001 2010-06-22 018[4]

What I did notice was that I had to get used to and adjust for the metering when using the PanoSweep.  Early on I ended up with a lot of shots that were blown out:

2010-06-22 002 2010-06-22 066

With some practice I learned to compensate by focusing on the sun first to set the metering, and then taking the sweep:

2010-06-22 002 2010-06-22 067

For comparison I also took multiple “normal” shots with the WX1 using the standard auto shooting mode.  The top photo of each is the WX1 PanoSweep version and the bottom version is a stitch composed from 4 shots using Windows Live Photo Gallery.  [Disclosure:  I work on the Windows Live Photo Gallery team at Microsoft.]

WX1 PanoSweep:

2010-06-12 001 2010-06-12 023

Photo Gallery stitch from 4 shots:

The pitch at Rustenburg before the US vs. England game

WX1 PanoSweep:

2010-06-12 001 2010-06-12 033

Photo Gallery stitch from 4 shots:

Rustenburg Stadium Panorama 2

Overall the Photo Gallery panoramic stitch clearly wins out on quality. This makes sense because it’s composed from four higher quality (~10MP) images (The WX1 down samples the image to 5.1 MP when shooting in PanoSweep mode).  That said the WX1 is far more convenient, and I remember to change the dial on my point and shoot more often than I think “let me take 5 images to stitch later.”   Having the result of the panoramic available immediately on the camera LCD is also encouraging because you just end up trying the shot a few times until it looks good.

The one question I got about the PanoSweep images was if they are cropped automatically.  The answer to that is both yes and no.  The image from the WX1 comes out cropped automatically, but that’s because if you veer too far from center vertically while capturing the panorama, the camera will spit out a message saying something to the effect of “Please move the camera steadily horizontally.”  I’m assuming the WX1 is actually capturing somewhat above and below what I see as the final shot, and only spits out the middle vertical section of what is captured as the “final image.”

You can also change the PanoSweep setting to expect the camera to sweep upward or downward.   Again, I assume the WX1 just has an accelerometer in the camera that keeps track of how close your sweeping motion is to center.  Also, if you’re hoping the WX1 will let you shoot an infinite panorama, you can’t capture indefinitely.  The camera senses a certain range of motion and once you’ve completed that range it spits out the panorama.  This probably helps keep the WX1 panoramic stitching within the processing capability of the camera.

MacHeist NanoBundle 2

I purchased the MacHeist NanoBundle this morning. I haven’t gotten much of a chance to look at the apps included yet, but a big draw was MacJournal, which appears to be compatible with independent wordpress installs (this is my test). Also Tracks looks like it could have some serious potential. It’s a little menu bar icon that allows you to search your itunes library and find/play a specific song super quickly. It also scrobbles to last.fm for you if you want (I do want). I’ll put up some more info as I have a chance to play with the applications more, but until then you’re probably best of going to macheist.com to look at the list of programs yourself because I believe there’s only one more day left to purchase.

Defamation 2.0

Cal Tech and MIT have had a fun rivalry going for a few years now (or rather the practice of hacking between the two schools increased during my undergrad years). Knowing that I wasn’t too surprised when I received a link to mitrejects.com and saw the end result.

It’s a relatively harmless and humorous little joke, but I wonder how the Cal Tech administration will respond. More interestingly, I wonder the Cal Tech administration can do. I doubt there is any law on the books anywhere that mentions “defamation via domain name.” My guess is that Cal Tech will look of the owner of the domain name and ask them nicely to stop redirecting it to caltech.edu. The joke may end there, but if the owner of the domain name politely declines the request, I expect that Cal Tech will look into legal options before ultimately realizing that the mostly likely option for resolving the issue will be to issue a request or “take-down notice” of some sort to the domain registrar for the domain mitrejects.com. At that point it would not at all surprise me if the domain redirect was killed by the registrar, trumping the wishes of the actual owner of the domain name.

It’s a realistic possibility, and it begs a couple questions:

  1. Should there be laws preventing this sort of internet bullying/pranking?
  2. Should your domain registrar really be able to force you to accept terms of use that cede your control of the domain to them?

Note: this is all entirely speculative, and potentially completely wrong and inaccurate. I however, don’t think it’s a bad guess.

The iPad’s One Flaw

While walking home this evening from the light rail stop I heard the most interesting criticism of the iPad anyone has mentioned yet:

It doesn’t even have a DVD player.

Listening to the concerns of people who don’t make their living around software or technology can sometimes be quite enlightening.

Take a moment to think before switching to the Google public DNS

Today Google has announced they’re running a domain name server free and open to the public.  The Google blog post talks about how Google feels compelled to improve the speed and quality of domain name servers because it will help make the internet a speedier and happier place for everyone.  That very well may be true, but I would assume there’s also another side to this not mentioned in the blog post.  While Google may offer a great DNS for you to route your visits ESPN.com, Amazon, or Facebook through, they’ll probably also now have a guaranteed way to know every single site and page you visit.  Google probably has lots of good ways to get most of this information from you already (through their existing services you probably use), but as a domain name server maps the name of every URL you type to the actual server hosting that site, this should guarantee they get all the information they want about your browsing habits. I can’t imagine that Google is not using this information to their advantage in some other way as well.

I’m sure there will be plenty of people that will immediately begin using the Google Public DNS immediately thinking it’s the greatest thing since slice bread, just like there is always a group of people that line up to buy the newest Apple gadget.  This may be the best domain name server created, but you may want to stop and think about what other benefits Google may be reaping by providing this service.

Privacy is not dead on the internet, it is however very easy to unknowingly sacrifice.

Random thought, would it at some point be good for the Google homepage graphic to be Google with the two Os replaced by binoculars?  Maybe that wouldn’t portray the right image Wink.

VMWare Fusion Unity and Windows Live Writer for awesome client blogging on my Macbook

So I know this setup is a stretch for most people, but it’s something I’m pretty happy about.  I’ve previously written about the fact that I think Windows Live Writer is awesome, and I still do.  In fact my recent revival on the site is largely due to the fact that I recently installed a fresh build of Writer at work to post and revise, as well as test and help weed out any bugs/failures.

Previously I’d said that if I could run Writer on my Mac to do my blogging, I would.  Well now I’ve made that statement a reality.  I’m running Windows 7 Ultimate virtualized under VMWare Fusion 2.  On my virtual machine I installed Windows Live Writer (Wave 3) and when I run the application in Unity mode it works almost exactly like a native Mac application.

It looks something like this:

Windows Live Writer on VMWare Fusion Unity

[update] I happened to forget the disclosure I had intended to include.  As I am currently employed by Microsoft, it’s likely that my opinion on anything involving software will be interpreted as biased.  While I do work on the same team as the people that make Writer and I know plenty of them, this site has absolutely nothing to do with my job or employer.  I’m under no obligation whatsoever to talk about software, I just happen to like Writer.  It’s a simple application with a well defined purpose which I believe it accomplishes quite well.

Conversation Trace

Since I’ve entered the new (to me) world of client development, I’ve had to re-learn a lot of very basic things, such as what some commands are.  At times it leads to unintended geek humor…

 

campo [11:19 AM]: is the command k in ntsd a stack trace?

tesh [11:19 AM]: yup                                                                            

campo [11:19 AM]: k

tesh [11:19 AM]: lol