MSN Search Coming on Biased?

Robert Scoble applauds the MSN search team for putting Microsoft.com search results in first place:

First off, did you notice that if you search for Windows API calls that MSN Search now takes you to the right place on MSDN? That's a huge improvement and one I gave MSN a lot of s**t about earlier. For instance, if you search on DestroyWindow it used to take you to my blog. It now takes you to the correct page on MSDN.

That's a HUGE improvement.

Whether microsoft.com search results should come up for these search queries - or not - should it really be Robert Scoble's decision? 

I wonder what things they had to do to the algorithmic search rules, index, whatever to make this happen?  What's next, I search for "operating system" and Mac and Linux doesn't come up until the third page?  Because Windows isn't on the first page right now, somebody should tell Robert to publicly lambast the "Team".

I am not saying that MSN search is biased. I am just saying that if it is an algorithmic search engine, let's let it be an algorithmic search engine.

New proposed tagline: MSN Search: Edited by Robert Scoble.

All that said, MSN Search is playing a good game of catch up.

Comments

[Randy Charles Morin]

Whatever they did worked for Win32 applications. That's just great, but who is developing to the Win32 API anymore? I thought we were doing .NET? And as I pointed out in Robert's comments, MSN is not in the game for finding .NET classes.

[Gabe]

Peter, the change Robert observed put Microsoft's results for "DestroyWindow" more in line with Google's and Yahoo's results. So you're assuming that MSN search achieved ends that appear more "algorithmic" through "biased" means? Why?

[Michael Fagan]

I'm going to have to (sort of) disagree with you. I think the best tools are those that combine algorithmic results with the knowledge of people. Of course, the algorithmic results are based on people too..

I'm working on a search engine for content related to my university, for example. If you search for "CS 100" (a course code) there will be a couple of links at the top, such as one to a listing for that class for the current semester. This is a link to a dynamic page (think 'invisible web') that wouldn't normally turn up in a search engine at all, but because I've 'hardcoded' it, it does. I think there's a lot of value in that. It'll also link to a course description, which is somepage.com/foo/#coursecode; i.e. a *part* of the page. That doesn't happen with any algorithmic search engine today that I know of.

[Robert Scoble]

I didn't say that they messed with the result set with humans. Their algorithm is being more finely tuned so it is giving better results.

By the way, Google's results for Windows API calls have pointed to the right page for years.

[peter caputa]

Yeah. This was a stupid post. Sorry for wasting everyone's time.

Thanks for the clarification, Robert.

------------------------
As always, Michael, interesting stuff you're doing.

[Michael Fagan]

No, not a waste of time.

I didn't stress the *other* side of the equation. Deciding what's a *better* result is always subjective. I'm sure all the major engines do plenty of quality testing with large groups of "average users."

But it can still be quite easy to say "our coverage of [event] is the best around, so we should put it first". Yahoo always puts their content first (marked with a 'Y') http://search.yahoo.com/search?p=travel . So there's a fine line between "improving the user experience" and pushing one's own content above others.