Back in the mists of time, a couple of Stanford PhD students called Larry and Sergey came up with an idea to make searching on the web better. This they undoubtedly did. But skulking in the low-rent dens of Silicon Valley and elsewhere an army of wannabe moguls work tirelessly on usurping the Mountain View giant.
Whether for niche searches or for the whole web search prize, whether by human-powered or visual methods, whether by developing cunning semantic algorithms or by Social means, all these platforms have one thing in common: they all want to take a bite out of Google.com's posterior.
Visual SearchTop Dog:
Quintura provides contextual visual search, and has long been touted as a genuine contender for the next-gen search king crown. You enter a search string or hover your mouse over a tag in the visually represented cloud. The cloud drills down and offers you more tags to hover over, and drills down again until you have reached the tag you want.
High-profile converts like ReadWriteWeb.com and the Maxim stable of websites may just bring Quintura into the mainstream.
Kartoo, a visual meta-search engine, has attracted more than its fair share of attention. Supposed to make search 'fun', or something, by lobbing in a Flash interface for your search map. Takes too long to load, but on the plus side boasts a feature-rich interface and page previews, not to mention a genie riding a surf-board.
'Search The Tail takes Google a step further by allowing users to narrow their search results by using either popular keywords or obscure terms. The keyword list, namely Google Tail, is in order of popularity.'
Whatever this is, it looks like nothing you've seen before. It's like Michelangelo and Steve Jobs got together to do visual search. If there was ever a website that's gonna help you to pick up chicks down at Starbucks just by looking at it, you've found it. So that's good enough for us then.
Apart from being stunning-looking, you can target search to stuff like stocks, news etc. Well worth a look.
Exalead is a French SE which provides thumbnail page previews of results, and shows various extra information on the results page including maps, file types, categories, rich media and so on. Apparently you can also use it to cheat at crosswords.
Human SearchTop Dog:
Mahalo is touted as 'The First Human-Powered Search Engine', and it's here to help. Set up by Weblogs co-founder Jason Calcanis last year, the Mahalo community filters out garbage links and provides you with quality pages and multimedia for your searches. Popular topics have their own 'Guide Notes' with all the essential information you are likely to need (eg here). Of all the current human-powered Web search alternatives, Mahalo is probably the most likely to succeed in the niche, given exceptional growth over the past year.Others:
Trexy is a meta search-engine which relies on the 'Search Trails' of users to streamline the search process. If someone has made a particular query before you, and had to wade through several resources to get to the information required, then that's information that is useful to you. This search tool provides you with this information, so you can in theory avoid the mistakes made by those coming before.
To see how the trails work, look at this query for 'how did Columbus discover America'.
Trexy generated a lot of hype soon after launch in 2006… and promptly fell off the radar.
Clustered search results. 'Instead of delivering millions of search results in one long list, our search engine groups similar results together into clusters. Clusters help you see your search results by topic so you can zero in on exactly what you’re looking for or discover unexpected relationships between items.'
True Knowledge's 'Answer Engine' aims to streamline the process of getting a specific answer to your question. It does this by 'structuring data in a way that enables computers to work and think like humans do, drawing inferences and conclusions when needed to find the information that's requested.'
This thing is in Private Beta, so we can't tell you how good it is. But you can already use its API to utilize the platform in your Web services. In fact there are already a small army of specialized API services for stuff like IP Geolocation and 'phone-number-to-location'.
Members essentially vote on documents which relate to a specific search term, so you should get good results. Unfortunately, Stumpedia's Web search has a long way to go — we tried 'Google history' and 'computer virus', and both returned the dreaded 'No Results Found' message.
But there's better news elsewhere. A recent new addition is the 'Instant Answers' service, which simply connects people to the Stumpedia community, with a view to answering queries. You use a sort-of IM client to connect with people, ask them a question, and wait for a reply. .
Xoost is one of the front-runners in the burgeoning 'social search' scene, which overlaps in a significant way with the 'human search' movement. You enter a search string as per normal, and get your regulation results. But… your experiences in finding relevant information are saved and shared with other users, and so the system learns how to hone in on better results. Go to Xoost to grab an invite to the Beta, if you're interested.
ChaCha invites you to ask any question in plain English from via Cell phone SMS or over the Web. Currently in Beta, ChaCha relies on human 'Guides' to answer questions, and then send the answer back to you via SMS/Web.
Semantic and Natural Language Search Engines
A Note on Semantic Search: Semantic search relies primarily on content being in RDF format, which is a particularly rich descriptive format. One drawback to Semantic search is that results take aaaaggggeeesss to get back to you. These things have to do a lot more drilling to get more accurate information from the deep recesses of the Web.
Microsoft recently opened up its fat wallet to buy Powerset, a search engine which enables natural-language queries. The move ended lengthy speculation surrounding the company's future, and put to bed persistent rumors that Google was willing to break the bank to acquire the start-up. VentureBeat explains the attraction:'Powerset, of San Francisco, has developed a technology that attempts to understand the full meanings of phrases you type in while searching, and it returns results based on that understanding. By buying Powerset, Microsoft is hoping to close the perceived quality gap with Google's search engine.'
At the moment, PowerSet only works its natural-language magic on Wikipedia articles. Like this.
Hakia is one of the most serious players in the Semantic Search space, and up to Dec '07 Hakia had raised $18m in funding. Hakia's quest for quality results hinges on three simultaneous criteria: that ” It (1) comes from credible sources (verticals) recommended by librarians, (2) is the most recent information available, and (3) is absolutely relevant to the query”. Services include Web search, embeddable search, Web services and an API are available. In public Beta.
Swoogle is a Semantic Search research project run out of the University of Maryland, and has a logo and name which is likely to attract some stern legal correspondence with a Santa Clara, CA, postmark if it ever goes commercial. As well as browser-based Web search, 3rd party Web services can utilize Swoogle's crawl.
Semantic search powered by RDF Gateway. And Web services with sample agents and a Web service interface to get you started.
- Endangered Species: The Corporate Intranet
- Think Digital Marketing Technology: Think ... Microsoft?
- Make Room for Gartner's BI and Analytics Platforms MQ Leaders
- Will Office 365 Destroy Consulting?
- Multitasking? You're Killing Yourself for Nothing
- Forget Intranets, Give Me an ESN
- From Build It and Go, to Ready to Go with SharePoint