Google's Test Search Engine

Bengt noted Google's SearchMash which is a testbed search engine. Google spokesbot says: "The goal of Searchmash is to test innovative user interfaces in order to continually improve the overall search experience for our users. The experimental search engine looks very different from Google's Web sites and lacks Google branding. In this way, Google believes the site will yield more objective feedback from users."

Comments

"google censorship"

[Wonko the Sane]

Try searching for "google censorship" and it shows:

Web Pages - about 190,000

And relevent wikipedia articles

I guess it works.

Re:"google censorship"

[Andrei D]

What it's odd here is that the wikipedia page is yielded twice: in the web pages section and in wikipedia. I think its common sense to display it only once.

Text browsers

[dattaway]

It doesn't work in Links.

Leechy

[ostehaps]

Nice. Just unfortunate that the results page looks uncannily like on of those domain leech default pages.

So how is this different from google.com?

[diegocgteleline.es]

It looks pretty much like google.com but with a different brand and logo. Even the colours in the result page are the same....

The one "new" thing are implicit image search in each search...and google was already doing that with some searchs

So how is this a "testbed search engine"? And why the article writes "the experimental search engine looks very different from Google's Web sites", when from a first look it clearly isn't?

Re:questionable reference built in

[Digicrat]

That's a very good question.

I also wonder if Google, aside from this, has donated anything to the wikipedia foundation? Google does claim they support open products, and Wikipedia has been at the top of most regular search results for a while. That doesn't give them any obligation of course, but would be a good-will-promoting-and-tax-reducing-act towards a non-profit that complements Google's offerings.

Of course the real controversy here should be that does including Wikipedia as a special-case in search results lead to a greater trust in the accuracy of Wikipedia's content? And is that trust merited?

Re:questionable reference built in

[Anonymous Coward]

This shouldn't be controversial. This is the aim of Wikipedia. Wikipedia is not based on commercial interest, but rather, collecting and synthesizing information to be used by whomever can make use of it. Google is using Wikipedia's information to help those who search.

Since its about user interfaces, why not a contest

[3seas]

That's right, a contest or better yet a ul/dl your customized search engine interface to google engines?
You know, like firefox has skins and other goodies the users create and share.
So how about an easy to use skin development package to the google APIs?

I like the idea of seeing samples of other searches, like images and groups though I might be doing a search on web.

I know google is about advertising for their income so somehow thats gonna need to happen.

Dragging Results

[Siker]

For a while you could also rearrange the search results by drag and drop. Last month I wrote a little conspiracy theory about the true purpose of all this dragging. Seems like they removed that feature now though, so I guess that's a sign I was wrong. Or maybe they saw my blog and realized the secret was getting out and hid the feature... :)

Re:feedback on the feedback

[Kijori]

I can't see any feedback form, but as someone mentioned, they review what people search for, so you could just search for "images box is too far down".

Problems with "++" searches

[macklin01]

Strangely enough, I had trouble searching for C++.

For example, C++ jpeg gets turned into C jpeg, and returns a bunch of C code. If you search with quotes, "C++" jpeg, you get "C " jpeg. Search for "devc++", and you get "devc " and information on Devcon international.

This doesn't make the search engine particularly useful for C++ coders. ;) -- Paul

Re:Horrible

[springbox]

I don't have a problem with sites that use JavaScript to do neat things, but it's dangerous to rely so heavily on it. I have used "web 2.0" sites that did an all JavaScript search on a single page.. After doing a few searches, I wanted to go back through my search history but pressing the back button took me to a completely different site (the one I was viewing before it.) Things can get like that if designers get too comfortable with their neat interface and can break a lot of things. Like, bookmarking the current page always takes you back to the site's "home page." The same thing happens with sites that present themselves only using Flash. Those are a pain to navigate and return to.

Re:I liked the more webpages feature

[robaal]

But only up to 100 results. I liked the MSN search beta more, where the results would be loaded automatically while you scrolled down; I don't think that one had a limit...