TechnoCloud

That man in constant motion Robert Scoble points to Quintura, an intriguing search engine based on refining your search through a tag cloud.

It boasts a clean interface, accompanied by a helpful one minute tour, and I found it simple to use. You type a query, related search terms appear in a tag cloud which you can add to your search by clicking on them, refining the search results below. You can roll your mouse over related search terms to see hints for additional search terms or click on an ‘x’ symbol to remove that term from your cloud.

It’s an interesting way of quicky refining your search results for ambiguous terms, e.g. spears (Britney or weapon?), something that Google and others have looked to address through predictive search ‘did you mean’ links within the search results.

The search results state that they are ‘Powered by Yahoo XML’ suggesting that Quintura technology can be built on top of other existing search services.

The service tries its best to include the relevant Web 2.0 must-haves, starting of course with the obligatory blog. However, ‘Share it’ at present just emails a link to Quintura to a friend rather than an extended social network of recommended sites. Likewise ‘Save it’ just shows you how to save a link to your search rather than any kind of social bookmarking. 

These elements of course can be developed over time and this is at present a ‘beta’ (what isn’t these days?). Their corporate site indicates there’s more plenty in the pipeline, so this is one to watch.

Adaptive Blue have released version 3.0 of the BlueOrganizer social bookmarking tool, with a Firefox extension that is a contextual search menu linking you into the ’social mediasphere’. On first look, it brings to mind the Hyperwords project with its contextual browser extension for interacting in more depth with the keywords on a page.

After a brief install and browser restart, you can right-mouse and see a ‘BlueMenu’ either for a highlighted piece of text or for a page as a whole. The menu opens into all the poster children of social media, with stalwarts Wikipedia, del.icio.us, Flickr and the rest just a menu click away.

In theory, this click takes you directly to the tag or results page relevant to your selected term, but it doesn’t always work in practice. It worked fine for me on del.icio.us and YouTube, but just took me to the homepage of Wikipedia and an ‘all tag’ page on Odeo for my chosen search of ‘media technology’.

Helpfully, it opens the page in a new tab allowing you to explore multiple choices without losing your original context. There are revenue models in place for the existing product, with paid placement possible in the menu of default options and affiliate links to take you through from your search for a film to buying the DVD.

Why are we waiting?

Ironically for a tool that is supposed to save you time, my biggest grievance is with the menu load times - on one occasion over 20 seconds and regularly more than 5 seconds. This needs urgent attention, if its to evolve from a bit of fun for early adopters into a genuinely useful tool. It appears I’m not alone in this (see comments).

Also the tutorials could be a lot simpler and shorter, if it’s going to gain a large audience. Better to focus on the basics and let users learn more as they become more familiar with it.

Certainly the idea of contextual search is an interesting one that has many potential applications, particularly when the user is able to customise their search menu, evolving it into a viable portable digital lifestyle aggregator.

Image 1: loading - you may be watching this for a while…

Image 2: the menu in action

It can be difficult to find photos you know you’ve taken. Most of us will by now have hundreds, if not thousands, of digital images strewn across hard drives. The more diligent among us will no doubt spend time carefully categorising and tagging each precious image. Some have even posted a power tip or two for speeding up this activity. For the rest of us, image analysis may help. Picassa should soon have image analysis (follow this link for more information.) Another potential solution is being developed that will be able to recognise elements in photos and videos by the Acemedia consortium.

Visual search company Riya have unveiled Like.com a visual search shopping portal at present limited to high fashion items which is the talk of the blogosphere. Riya.com itself was impressive enough with its facial recognition technology, but Like.com shows several improvements. Users can search among the current inventory of jewellery, handbags, shoes and watches and through a ‘likeness search’ can query the visual search engine for products with a visual likeness.

Like.com gives searches the ability to find related objects of a specific colour, as well as by shape and pattern and rank their relative importance through sliders.

Coming soon will be the ability to upload an image, e.g. camera phone shot of an object you like, and then find its likeness, with the current example on the site being finding similar items to those celebs are wearing.

Furthermore, the technology allows you to zoom in on a particular characteristic of an item, e.g. a high heel on a shoe, and narrow down your search to items that share that characteristic. This particular functionality has wobbled here at TechnoCloud with some loading problems, but we can excuse that as alpha launch blues.

All in all, an impressive piece of technology, if not quite matched by the early doors limited product categories. The real value for this may lie in licensing the technology to third parties to allow them to improve search facilities and categorisation across a range of industries, from clip libraries to mapping/tourism (e.g. what building is that?) to myriad different shopping applications.

Visual recognition is one of the significant challenges facing search as the scale of multimedia content on the web grows. It is relatively straightforward to extract text and even audio, but the variables involved in visual recognition mean that a human eye remains the best judge of a successful search. The likes of Blinkx, Riya and eVision are trying to change that but it remains early days.

Interesting then to see the Retrievr experimental service. It allows you to draw a Paint-style sketch or upload an image and then to search a selected group of ‘most interesting’ Flickr images for matches. Rather than object recognition, e.g. find ‘chairs’, it works by searching for related blocks of colour and overall shapes (see pretty accurate results for a beach below).

Although the existing link is just something to play with at this stage and the results are mixed, the technology has powerful potential to retrieve better search results. For example, rather than just a search for a dog, you could use this in conjunction with existing search to find a dog on a blue background that fills the majority of the screen.

This could have all kinds of potential applications, such as better tagging of visual archives, finding unauthorised use of your copyrighted images, finding a particular video clip.