Search Wikia At Last - The Power of People
If you're new here, you may want to subscribe to my RSS feed or to sign up for email alerts.
Do you ever get tired of downright stupidity? I know you do and I sympathize whole heartedly. In the technology field of writing we often wrestle with complex and difficult to explain innovations and services. Some of us spend untold hours just trying to understand, let alone explain some of these creations, and even then we fall short in explaining them - and we feel we fail. Then there are people who do not even bother to do the research, and they offer us what I would call “the Walmart” version of a story. You know the ones I mean, those that do not know what the hell they are talking about.
The other day a colleague of mine, and I will not say, offered a simplification and downright insulting article about Search Wikia. The search offering by my friend Jimmy Wales had just gone into Alpha testing and many people covered the story. This version compared Search Wikia with Mahalo, which is like comparing grits with caviar, and the piece essentially ignored anything that has ever been written about the fledgling engine. Is it not our job to reflect to the best of our ability the potential and capabilities of these technologies?
What Is Search Wikia Really?
I have written about Search Wikia several times and one thing it is not is a recommendation engine or Mahalo. The original vision of Wales and his collaborators was to have an open discussion and development for a semantic engine that is refined by human scrutiny. I expect that the initial plan was fairly rough but it might have included some of the same technologies that hakia and Powerset do only on a smaller scale. The backbone of the engine was and is the human capability of selection. The following is an excerpt from one of my earlier investigations of Search Wikia.
“To get the semantic meaning of the website the crawler also uses the WikiRelate! algorithm (or something similar) to link the keyword to a wikipedia article trying to do a best-fit mapping of the website with a URL of the type http://
.wikipedia.org/wiki/ , where the xx-Part is extracted using the xml:lang Attribute. If the lang-Attribute is not set then we try a mapping using all Wikipedia Articles from all languages. Alternatively we might implement a new algorithm that uses OmegaWiki’s database for guessing the right language.”
As you can see the early development of Search Wikia was a forum for ideas to develop an engine that encompasses all modern thought with regard to new age search. Does anything about the passage sound like Mahalo? This excerpt was taken from the very first entries by contributors to the Search Wikia effort. If my good colleague had taken the time to visit Search Wikia then I think a more correct and appropriate dissemination of the facts of the new ALPHA could have been rendered. The basic principles of Search Wikia are: transparency, community, quality and privacy. That being said, it is also easy to see that this engine is not intended to be any sort of suggestion engine or an excessive people filter.
The active areas of focus are: “taken from the site”
Social Lab - sources for URL social reputation, experiments in wiki-style social ranking.
Distributed Lab - projects focused on distributed computing, crawling, and indexing. Grub!
Semantic Lab - Natural Language Processing, Text Categorization.
Standards Lab - formats and protocols to build interoperable search technologies.
The term interoperable stands out here for a good reason. Search Wikia was never envisioned as a simple index of pages submitted or recommended by collaborators, but a rather more complex integration or human minds, algorithms, and yes perhaps even semantic interoperability. In the end Wales and his collaborators envision essentially the same thing that Barney Pell of Powerset and Riza Berkan of hakia do - better relevance. The ends to by which the goal is met are irrelevant to an extent, but the complexity of these engines far exceeds anything Mahalo or others might be doing .
Conclusion
It is important to note, as my most trusted colleague Mihaela Lica did in the comments of the aforesaid article, that Search Wikia is still just in Alpha testing and in full view since the early days. I know that Search Wikia deserves more than a cursory glance and a comparison to Mahalo, as it has been one of the most anticipated search entities. I am always appalled at just how quickly people are willing to negatively scrutinize potentially great technology. I for one am rooting for all these people and friends so that we might see relevance light years beyond what is currently available, regardless of the methodology used to get there.
It’s a good idea, but open to abuse. Do we really need another search engine?
Okay, I may be one of those who don’t know what the hell I am talking about
This is from their website:
We are aware that the quality of the search results is low …
Lu
As mentioned, Search Wikia is still in its infancy stage in Aplha. Many are quick to judge without giving things a fair deal before giving them the time to evolve and improve.
Luci, the quality of the results is still low because they are waiting for the people to join in and refine the results. It is a collaborative project, as everything else “wikia” is.
It will take time, but let time tell how powerful people really are.
Hi Phil,
Looks like the Wikia approach involves building a whole new community in order to build up the relevance of search results. Instead of starting on a clean slate, how about leveraging on the current leading social bookmarking communities (e.g. del.icio.us) to enhance the combined metasearch results from the major search engines (Google, Yahoo! Search, Windows Live Search)?
We’ve just launched a web search service called
FuzzFind (http://www.fuzzfind.com) that does this. Results are grouped together and sorted according to the search engine rankings plus the popularity of the sites according to the social bookmarking community.
There’s also a feature to allows you to tune and personalize the search results on-the-fly by simply adjusting the weight for each search source through the use of sliders.
Hope that you have some time to check it out, any feedback is welcome so that the service can be improved. Thanks!
Thanks, Phil!
The main thing is that this is a project to build a search engine, not a fully developed search engine yet. There’s a lot of hard work to be done.
Mahalo is a human powered search directory that uses actual human editors to compile results alongside Google-powered results.
Search Wikia is a semantic engine like Hakia and Powerset, only on a smaller scale.
Semantic search engines use semantics - the science of meaning in language - to produce relevant search results. rather than ranking algorithms such as Google’s PageRank to predict relevancy, This requires that is all the knowledge used by the system, is represented in the form of a semantic network, organized on a conceptual basis. [Semantic search - Wikipedia] In Wikia this system is based upon Wikipedia
It is also clear that comparing Mahalo and Search Wikia is indeed like “comparing grits with caviar”, moreover as stated explicitly by the developers Search Wikia is first alpha release and “[they] are aware that the quality of the search results is low”
Let us - instead of criticizing it with irrelevant or unfair objections - follow with an open mind the further development of the fascinating Wikia initiative.
@ Deborah,
Exactly gal - it never ceases to amaze me how people miss the fact that these early versions are simply previews (and often very rough ones) of the developments to come.
Jimmy has always turned out fantastic value and innovation - so what makes anyone think this will be any different?
Thanks for the comment Deborah!
Always,
Phil B.
Thank you Rob - Objectivity and also the track record of Wikia seems relevant here. Even if this were a completely new fledgling company - people should be rooting for innovation rather than condemning it.
Search Wikia will work in my mind, and I don’t really see it failing in the long term. Like the services before it from Wales, it will occupy a niche of its own and be of great use to many people. As for beating Google, I am not really sure that is the purpose, but rather simnply adding value to the Web.
Thanks,
Phil B.
Hi Fuzz! I could not agree more with your idea of integrating other data segments into the engine. I will also be apply to test your innovation at my earliest convenience.
Thanks,
Phil
Thanks so much Jimmy! I know you are covered up as always. I hope we can all contribute and see the end of poor relevance this year or at least soon.
It is very difficult to project the vision as you know. You have my support always and my great regard as well.
Always,
Phil
Hi Jimmy,
Are there any immediate plans for Wikia Search to provide an API? There’s mention of it in one of the principles for the search engine, I’m wondering if it will be sooner rather than later.
It would be great if we could combine the strengths of Wikia Search together with the rest of the major search engines and social bookmarking sites into our new service FuzzFind Web Search (http://www.fuzzfind.com)
Thanks!
Hey Fuzz… I cannot help a comment - I checked and tested your search tool and I’d like to say that you still have a lot of work there to make it competitive with at least some of the meta search engines (like Clusty Search for example)
It’s good that you’ve integrated social bookmarking, but del.icio.un is almost a dinosaur and if they don’t do something fast they’ll be extinct. Maybe you should consider including digg or Stumble?
Also, when you show results from the social bookmarking sites, why don’t you provide the link to the result, so that we all can see what the users were actually writing there?
As for search wikia, if you want to contribute, just follow the link and start working with the others. This is a collaborative project as everything Wikia - and I am sure your expertise would be appreciated.
Mig