I'm going through a little discovery here. I just typed code search into my FireFox location bar instead instead of Google, where I meant to type it. When I hit Return I found myself at the Google Code Search page: http://google.com/codesearch. Why did that happen? At first I thought it had to do with my having a Google toolbar in the browser, but then I tried it on Firefox on three different machines: Linux, Mac and Windows. Same result. So it appears that FireFox by default treats the location bar as a Google Search bar.
And indeed, I discover (through a Google search, of course), that "location bar search" is defaulted in Firefox 2.x. Why? (Pretty complicated to defeat it too.)
When I type code search into Google instead of Firefox's location bar, it suggests I should be looking for source code search, where the lucky top result is for Koders. Why wouldn't the results from the location bar go straight to Koders, then? The answer is that I did search for just code search (however unintentionally), and Google Code Search is the top result for that. But still, it raises suspicions for me. Why would it go to one of Google's own services?
Here's what's also weird. I just checked the Advanced Search page to see how many inbound links go to each of the code search engines I know about. In alphabetical order, they are AllTheCode, Codase, Gonzui, Google Code Search, Koders and Krugle. Here they are in a table, with rows ordered by inbound links to each one:
| Engine | Inbound links to the engine | Placement in search results for source code search | Placementn in search results for code search |
| Koders | 5,770 | 1 | 3, 5 |
| Krugle | 3,530 | 10 | (15th page) |
| Google Code Search | 1,580 | 2 | 1 |
| Codase | 327 | 3 | 4, 11, 44 |
| Gonzui | 83 | 8 | 51 |
| AllTheCode | 64 |
In addition to inbound links, I looked at placement in search results for source code search and code search. I started with the latter, because that's what I was looking for in the first place. The results for that are strange, though. There is one sponsored link, on the right, for Google Code Search. Then the results are divided into the groups separated by two horizontal rules. At the top are three results. First, Google Code Search; then an indented result under that for Google's advanced code search, then a result for O'Reilly Labs Code Search for looking through O'Reilly books. Then there is a horizontal rule, below which it says "See results for: source code search". Below that are Koders, Google Code Search and then Codase. Then there is another horizontal rule, followed by Google Code Search, Google's Development Network, SAT Code Lists by collegeboard.com, and finally U.S. Code Search at uscode.house.gov. That's it for the first page of results.
I'm wondering about two things here. First is why I'm getting search results in my location bar — especially one that appears to favor Google's own offering. Second is why the search results look the way they do.
What stands out for me is the results for Krugle. A disclosure: I'm an advisor to Krugle. The label at this point is a something of a stretch, however, because I haven't been in contact anybody from that company in many months. My search for code search, which precipitated this exploration, had nothing to do with Krugle. (It had to do with VRM, actually. The project needs some code.)
Anyway, I began to wonder why Krugle did so poorly in the code search search. As I go through the results I find...
When I look at all the Good Work that Google does, it amazes me. When you consider the size of the Web's vast haystack, you have to respect Google for finding needles in it. The company deals entirely in tall orders, and handles them with peerless abplomb.
But the company has also created enormous dependencies. For those of us who live on the Web, doing without Google has become a nearly unthinkable prospect. I am sure Google takes the responsibility that comes with this dependency very seriouisly. I am also sure that Google's anchor service — plain old Web search — has not advanced very much in the past few years. Certainly not in terms of (what I hate to call) user experience.
I just started looking through my Web History for Interesting Stuff. There is some, for sure. Like seeing that my daily search activity peaks on mondays at 1214 searches per day. Also that my hourly search profile looks like this:
Now, how many of those searches actually get me what I want? And how quickly?
What if I want to find, say, nothing but fly fishing sites in Idaho? How about if I only want results from sites last updated in 2005?