"Aylay is built on the collective intelligence of the Web. With Aylay Agile Search, start searching and get results simultaneously from the sites you know and trust, instantly. Refine your search and get results in real time without leaving the page."
That's quite a bit of marketese. None of that really tells me what it does. Is it a search engine? Does it search for trending topics? Is it a news aggregator?
I did a search for BioShock 2 and that showed me what it did. I think some better home copy would be "See search results from Google, Wikipedia, Twitter, Digg and more, all on one page." or maybe "Search multiple sites and see their results all on one page."
It's actually a pretty cool service though. I don't like the custom scrollbars on the results though. They just aren't standard enough. Maybe add a slider type function, or just a normal scrollbar.
I find it disappointing in its current form. I assume you aren't doing much more than a few ajaxified api calls to the 5 or so engines you're aggregating which is neither technically challenging nor technically interesting.
However, if you applied relative rankings for the individual items from each engine and aggregated them into a single results page instead of multiple "framed" result sections, you might have something I would actually want to use. This would also overcome many of the usability and UI problems I find issue with.
Because ~75% of clicks on search results are the top 5 results, we were less interested in the long tail of results, and focused on the top results from the most popular sites.
We developed this experiment to see if the top 5 results in context of the top sites would give a better result set than the top 15-20 results from any single search engine.
Well unfortunately, they just don't. I too have helped built a limited domain search engine. The site, www.trailbehind.com, crawls about 20 of the best hiking related sites, including wikipedia, everytrail, localhikes, etc.
Unfortunately, the long tail is where all the action is, and my site and your site will both be marginal compared to Google.
One way to think about it is they can always do what you do, but they can also use the context of the link graph that you don't have, so you really cannot ever hope to compare with what they do.
Everything you/I can do with 20 websites, Google could do if they wanted. But it's just not worth it.
Ideally you should develop a "velocity" kind of ranking similar to http://www.oursignal.com and merge in results, instead of simply displaying them in individual boxes. That way a user can work out which is the most popular/relevant entry overall, not the most relevant per site.
Thanks. I searched for one of my sites, and found a "how to hack [site name]" video on YouTube I didn't know was there. It wasn't actually a hack, just a silly JavaScript trick to disable a upgrade nag graphic, but I never would've noticed that video otherwise.
12 comments
[ 3.7 ms ] story [ 62.8 ms ] threadThat's quite a bit of marketese. None of that really tells me what it does. Is it a search engine? Does it search for trending topics? Is it a news aggregator?
I did a search for BioShock 2 and that showed me what it did. I think some better home copy would be "See search results from Google, Wikipedia, Twitter, Digg and more, all on one page." or maybe "Search multiple sites and see their results all on one page."
It's actually a pretty cool service though. I don't like the custom scrollbars on the results though. They just aren't standard enough. Maybe add a slider type function, or just a normal scrollbar.
glad you like the service, and we welcome any more feedback
a. 20% was injected between my words
b. I searched for "Income of women in India" and got the wikipedia entry for the United States.
I like being able to decide which search I want to look at however - very cool.
However, if you applied relative rankings for the individual items from each engine and aggregated them into a single results page instead of multiple "framed" result sections, you might have something I would actually want to use. This would also overcome many of the usability and UI problems I find issue with.
It certainly has potential to become something.
appreciate the feedback about consolidating results with relative rankings, that is something we are currently developing
Please tell me one search engine that doesn't search these. It's almost like you are calling out a weakness in your product.
Google doesn't search Wikipedia and Twitter and more. They search everything, and if you are going to be in the search game, you better follow suit.
We developed this experiment to see if the top 5 results in context of the top sites would give a better result set than the top 15-20 results from any single search engine.
Unfortunately, the long tail is where all the action is, and my site and your site will both be marginal compared to Google.
One way to think about it is they can always do what you do, but they can also use the context of the link graph that you don't have, so you really cannot ever hope to compare with what they do.
Everything you/I can do with 20 websites, Google could do if they wanted. But it's just not worth it.