Pikimal Blog

We like writing about all aspects of search on this blog, so a few weeks ago we decided to add a new category of post. Periodically, we’ll profile a small search engine that’s doing something different and interesting. You’ve probably never heard of them, but you might never forget them once you’ve tried a new way of searching.

The spotlight for this post is on Soovle. It’s a visual portal of sorts that searches all of the major engines at once. We asked its creator Matt Amacker some questions about Soovle and search in general. Learn all about it below.

 

Screenshot of Soovle.com

 

What inspired you to create Soovle.com?

I and one other engineer built the original search completion or search assistance service for a large website. When we built it, I was struck by two things. The first was how amazingly efficient it was. We could handle ALL of the typing traffic of every customer using the site on a single box with response times in the sub 10ms time. Of course, these were big boxes, and it wouldn’t run on a single box because we have to have industry strength redundancy and such – but it was still impressive. The second, and more important discovery, was how perfect the data coming out of it was for shopping. Many folks, at the time I was building this, said the data would end up looking like common web search because of the volume of queries and the tendency of popular memes to bleed across tasks. Not so! The data coming out of our search completion was so far superior to general web search that it left a mark. Which got me thinking…if ours is so different from the general web, perhaps there are others that might know their verticals better too, and given the open nature of Ajax APIs, the obvious low-cost of search completion responses - Soovle was born.

And where does that name come from?

There aren’t that many domains under seven characters available and I wanted something that could sound like a verb. Google had provided the training for how to say “oo” names, so Soovle was an easy choice.

Soovle has some neat tricks (downloading a csv, dragging search results, listing of top Internet keywords), why include these?

If you know engineers – they’re thrilled by making stuff useful and figuring stuff out. I’m no different – Soovle is the result of my play, it costs very little to support, and doesn’t make any money - but its value as my online jungle-gym is enormous. Some of the more
engineer-thrilling, yet undiscovered features:

• User’s can run Soovle with their own logo – with a link to their own site.
• User’s can “search together” – simply by passing a permalink to someone else, Soovle becomes a “shared” search experience – letting you see what’s typed and opening a chat window.

The thing that caught my eye about Soovle was how visual and fun it is. Was adding a more interactive and visual element to search something you initially wanted to do?

Yup – I make new and novel interfaces for a living ( scalable too! ). I was criticized for not making it a more “efficient” UI – but the rotating engines are so fun – I can’t bring myself to make it efficient. If you get a chance – head into an Apple store and bring it up on one of their Cinema displays, switch to 15 engines and hit the “demo” link. Hah – I love it! I guess I’m a little silly for being entertained this way – but I enjoy it so much.

Is Soovle your ideal search environment? If not, could you describe what a perfect search engine would do for you?

Soovle is not my ideal for all things. Its good for what people have discovered it for – digging up search terms from expert-equivalent systems for verticals. Google’s interface is far superior for general tasks and Amazon is far superior for shopping. So what I’d like to see is something like the Soovle engine rotations, on a Google search box, that lets me use Google for general search and allows the other sites to offer up their interpretation of my intent. If their right, it’d be great if I could jump over in a frictionless way. Of course, I’m an engineer so having busy interfaces is “ok” for me – this sort of interface is not likely to win over the “general web user”.

What is your tech-background? Working on anything exciting these days?

I’m an engineer for a big website – which means I get access to AWS to build cool stuff on top of robust systems – I’m building all kinds of neat things every day. Really, I have an awesome gig. I’m playing with everything from NodeJS to fascinating mobile technologies to the possibilities offered by the Kinect. I built Peerbind.com – which enables engineers like me to explore and play with peer-to-peer interactions. I have another service I’m working on releasing that will enable even more play – and I’m pushing the boundary of what people think the web should do.

Do you see anything wrong with Internet search as it is now?

Yes – specifically in internet search – we are missing too many of the ”cognitive ease” features that could make internet search more fun. One of the things about search completion and search suggestion is that it takes advantage of the brain’s fantastic “recognition engine”. We humans are able to recognize what we mean far faster than we can verbalize ideas into a screen. For instance – say you want to buy a ”car restraint kind-of-a thingy for my dog” so I start typing my guess: “car dog” then POOF! there in the completion is “dog car seat”, ”dog car crate”, “dog car harness”. With that I can immediately recognize my intent. Another way sites could assist me, is to provide more images, our brains are hardwired to do an excellent job of recognizing the results we are after. When I search for things that can be confirmed or sorted with images – they should show up and own a large portion of the page.

How do you see search developing in the future? What sort of features would you like to see added/embraced by the community?

I see search being more personalized – search suggestions that are tailored to me, and my interests, search results that make more sense for me. When I search for DynamoDB – I’m not looking for information about it – I need the API documentation, and my search engine will know that and save me time ( it already taught me about DynamoDB a month ago, and I’m beyond that now.