Today's big tech news is the launch of Cuil, the latest challenger to Google's hegemony in Web search. Given the impressive team of Xooglers that put it together, I had high expectations for the launch.
My overall reaction: not bad, but not good enough to take seriously as a challenge to Google. They may be "The World's Biggest Search Engine" based on the number of pages indexed, but they return zero results for a number of queries where Google does just fine, including noisy channel blog (compare to Google). But I'm not taking it personally--after all, their own site doesn't show up when you search for their name (again, compare to Google). As for their interface features (column display, explore by category, query suggestions), they're fine, but neither the concepts nor the quality of their implementation strike me as revolutionary.
Perhaps I'm expecting too much on day 1. But they're not just trying to beat Gigablast; they're trying to beat Google, and they surely expected to get lots of critical attention the moment they launched. Regardless of the improvements they've made in indexing, they clearly need to do more work on their crawler. It's hard to judge the quality of results when it's clear that at least some of the problem is that the most relevant documents simply aren't in their index. I'm also surprised to not see Wikipedia documents showing up much for my searches--particularly for searches when I'm quite sure the most relevant document is in Wikipedia. Again, it's hard to tell if this is an indexing or results quality issue.
I wish them luck--I speak for many in my desire to see Google face worthy competition in web search.
Subscribe to:
Post Comments (Atom)
Monday, July 28, 2008
Not as Cuil as I Expected
Today's big tech news is the launch of Cuil, the latest challenger to Google's hegemony in Web search. Given the impressive team of Xooglers that put it together, I had high expectations for the launch.
My overall reaction: not bad, but not good enough to take seriously as a challenge to Google. They may be "The World's Biggest Search Engine" based on the number of pages indexed, but they return zero results for a number of queries where Google does just fine, including noisy channel blog (compare to Google). But I'm not taking it personally--after all, their own site doesn't show up when you search for their name (again, compare to Google). As for their interface features (column display, explore by category, query suggestions), they're fine, but neither the concepts nor the quality of their implementation strike me as revolutionary.
Perhaps I'm expecting too much on day 1. But they're not just trying to beat Gigablast; they're trying to beat Google, and they surely expected to get lots of critical attention the moment they launched. Regardless of the improvements they've made in indexing, they clearly need to do more work on their crawler. It's hard to judge the quality of results when it's clear that at least some of the problem is that the most relevant documents simply aren't in their index. I'm also surprised to not see Wikipedia documents showing up much for my searches--particularly for searches when I'm quite sure the most relevant document is in Wikipedia. Again, it's hard to tell if this is an indexing or results quality issue.
I wish them luck--I speak for many in my desire to see Google face worthy competition in web search.
My overall reaction: not bad, but not good enough to take seriously as a challenge to Google. They may be "The World's Biggest Search Engine" based on the number of pages indexed, but they return zero results for a number of queries where Google does just fine, including noisy channel blog (compare to Google). But I'm not taking it personally--after all, their own site doesn't show up when you search for their name (again, compare to Google). As for their interface features (column display, explore by category, query suggestions), they're fine, but neither the concepts nor the quality of their implementation strike me as revolutionary.
Perhaps I'm expecting too much on day 1. But they're not just trying to beat Gigablast; they're trying to beat Google, and they surely expected to get lots of critical attention the moment they launched. Regardless of the improvements they've made in indexing, they clearly need to do more work on their crawler. It's hard to judge the quality of results when it's clear that at least some of the problem is that the most relevant documents simply aren't in their index. I'm also surprised to not see Wikipedia documents showing up much for my searches--particularly for searches when I'm quite sure the most relevant document is in Wikipedia. Again, it's hard to tell if this is an indexing or results quality issue.
I wish them luck--I speak for many in my desire to see Google face worthy competition in web search.
2 comments:
-
Daniel Tunkelang said...
-
I noticed that they at least fixed the search for Cuil so that their site now comes back. Some of my other searches that returned no results before are now returning good results, but it's a mixed bag. Perhaps they are learning from their logs. If they manage to rapidly improve their quality as a result of log mining or any other strategy, I will be truly impressed.
-
August 4, 2008 at 11:21 AM
-
Anonymous said...
-
At first I was impressed with the combination of images and text on Cuil. My assumption was they were picking the most relevant image from each result page, which seems like a nice idea.
Then I noticed that the images very often do not come from the result page, and sometimes do not match the text at all. And if you go through multiple pages of results, you often see the same images repeated. Has anyone else noticed this? Does anyone know what they are trying to do? -
August 5, 2008 at 9:11 PM
Subscribe to:
Post Comments (Atom)
Monday, July 28, 2008
Not as Cuil as I Expected
Today's big tech news is the launch of Cuil, the latest challenger to Google's hegemony in Web search. Given the impressive team of Xooglers that put it together, I had high expectations for the launch.
My overall reaction: not bad, but not good enough to take seriously as a challenge to Google. They may be "The World's Biggest Search Engine" based on the number of pages indexed, but they return zero results for a number of queries where Google does just fine, including noisy channel blog (compare to Google). But I'm not taking it personally--after all, their own site doesn't show up when you search for their name (again, compare to Google). As for their interface features (column display, explore by category, query suggestions), they're fine, but neither the concepts nor the quality of their implementation strike me as revolutionary.
Perhaps I'm expecting too much on day 1. But they're not just trying to beat Gigablast; they're trying to beat Google, and they surely expected to get lots of critical attention the moment they launched. Regardless of the improvements they've made in indexing, they clearly need to do more work on their crawler. It's hard to judge the quality of results when it's clear that at least some of the problem is that the most relevant documents simply aren't in their index. I'm also surprised to not see Wikipedia documents showing up much for my searches--particularly for searches when I'm quite sure the most relevant document is in Wikipedia. Again, it's hard to tell if this is an indexing or results quality issue.
I wish them luck--I speak for many in my desire to see Google face worthy competition in web search.
My overall reaction: not bad, but not good enough to take seriously as a challenge to Google. They may be "The World's Biggest Search Engine" based on the number of pages indexed, but they return zero results for a number of queries where Google does just fine, including noisy channel blog (compare to Google). But I'm not taking it personally--after all, their own site doesn't show up when you search for their name (again, compare to Google). As for their interface features (column display, explore by category, query suggestions), they're fine, but neither the concepts nor the quality of their implementation strike me as revolutionary.
Perhaps I'm expecting too much on day 1. But they're not just trying to beat Gigablast; they're trying to beat Google, and they surely expected to get lots of critical attention the moment they launched. Regardless of the improvements they've made in indexing, they clearly need to do more work on their crawler. It's hard to judge the quality of results when it's clear that at least some of the problem is that the most relevant documents simply aren't in their index. I'm also surprised to not see Wikipedia documents showing up much for my searches--particularly for searches when I'm quite sure the most relevant document is in Wikipedia. Again, it's hard to tell if this is an indexing or results quality issue.
I wish them luck--I speak for many in my desire to see Google face worthy competition in web search.
2 comments:
-
Daniel Tunkelang said...
-
I noticed that they at least fixed the search for Cuil so that their site now comes back. Some of my other searches that returned no results before are now returning good results, but it's a mixed bag. Perhaps they are learning from their logs. If they manage to rapidly improve their quality as a result of log mining or any other strategy, I will be truly impressed.
-
August 4, 2008 at 11:21 AM
-
Anonymous said...
-
At first I was impressed with the combination of images and text on Cuil. My assumption was they were picking the most relevant image from each result page, which seems like a nice idea.
Then I noticed that the images very often do not come from the result page, and sometimes do not match the text at all. And if you go through multiple pages of results, you often see the same images repeated. Has anyone else noticed this? Does anyone know what they are trying to do? -
August 5, 2008 at 9:11 PM
Subscribe to:
Post Comments (Atom)
2 comments:
I noticed that they at least fixed the search for Cuil so that their site now comes back. Some of my other searches that returned no results before are now returning good results, but it's a mixed bag. Perhaps they are learning from their logs. If they manage to rapidly improve their quality as a result of log mining or any other strategy, I will be truly impressed.
At first I was impressed with the combination of images and text on Cuil. My assumption was they were picking the most relevant image from each result page, which seems like a nice idea.
Then I noticed that the images very often do not come from the result page, and sometimes do not match the text at all. And if you go through multiple pages of results, you often see the same images repeated. Has anyone else noticed this? Does anyone know what they are trying to do?
Post a Comment