Google Hasn’t Improved SearchPosted: March 23, 2007
I’m getting tired of the Google love. There is without doubt no company I’d rather work for in the world – but increasingly because I’d like to give them a kick up the arse. Whenever they release a new product, it does nothing to improve the existing search offering. Whenever they do something to change the existing search offering, it’s a minor layout move. Whenever there’s a new product in labs, it’s no longer outlandish, it doesn’t make me think and again; is no improvement or change to their core offering: SEARCH.
Google hasn’t changed the way their search works at the core level in years, many years. It’s old hat and though it works well, they have the potential to make it so much better. They have a reputation for letting their developer teams think outside of the box, spending 20% of their time on personal projects.
When is someone going to go: “we need to move forward now, we have a wall full of ideas”, literally hundreds if not thousands (there must be at least one for each PhD) and it’s time to execute them in order to make Google a company not only committed to indexing all the world’s data but to letting us find it again.
I very much hope Google has something they’re sitting on, but again I don’t, because I’m not the only developer in the world working on technology in this area and as time goes on it becomes more and more likely they’ll be overtaken by an overnight hit from some small company in a small country.
I appeal to Google to return to “release early, release often”.
You’ll have to forgive the slightly random thought process occuring in this post, but I’m using the writing of said post to somewhat clear my extensively muddy thought wagon today.
Those of you who read frequently will know that I’ve developed a new way of analysing meaning in web pages or blocks of text. In short, using statistics I can now find words related to eachother in just about any language by studying how frequently they occur together on the internet.
Using this technology I have another idea for its usage that changes the way we order information on the internet. Essentially it removes the need for users to categorise their posts in any way, and is the natural progression of tagging.
Simply by setting up a website where people can post in one click their thoughts on a particular subject, the weather today or practically any thoughts that come into their head we can create a large index of information that sequentially is practically useless and has no context to a viewing user. We then take this information and present it with a semantic search engine, where all the returned results have been analysed for their semantic relevance to the query.
It enables us to have a wiki with no categories, to analyse otherwise unimportant data and compile it into a relevant list on any number of topics or any number of uses. This has tremendous power for existing sources available on the internet now.
Let’s apply this to an existing web phenomenen: Twitter, which is going from strength to strength presents the lives of individual users in bit form with constant updates in short sentences and descriptions. This is a perfect data source for this kind of project and would enable users to search for people with similar interests (think dating?)To find completely unbiased commentary on current events, products and ideas.
This is just another example of what the semantic layer I’ve developed is going to be able to achieve and I’m happy to say that we hope to move into a real Alpha release for Allegro (the desktop search engine) very very soon. The download site is being set up, we’ve decided to release under the GPL as a completely open source application and are really hoping to inspire the developer community to put together applications like this that really will change the way that people use the internet not only to find information but to gather it as well.
We’ve recently being doing a lot of work on the GUI and the way we present data, for those of you who checked out the youtube video we’ve come a long way since and it’s extremely slick and intuiitive. One final bit of news on that front is we’re going to support GUI plugins so that you can change the way you get your results, and you’ll also be able to pull results from Google if you don’t want to compile your own index.
This is improving search, this is taking data from every available source and regardless of content making it potentially useful, and making it easily locatable. Google, you have so much more opportunity, server space and money than I – so show us you still care. “Do no evil”, and “Do nothing, in case it’s evil”, are not the same philosophies. If you have something that you can contribute Mr Brin and Mr Page, that will further help the world, sitting on it is evil.
As always you can reach me at firstname.lastname@example.org all weekend, and then email@example.com when I get back to the office. Please leave comments with your thoughts, criticisms and plans for world domination.