I could see the need for relevancy in search even at that time. The speech might seem a little dated but the piece that is not dated are the statistics on the growth of knowledge in general. This exponential growth also applies to knowledge bits in the enterprise whether they be documents, blog posts, IM threads or just enterprise facts stuck in the head of a knowledge worker.

We can manage these types of knowledge in their own taxonomies that are best suited for each type but in the end the power to search for these, either by full text relavancy, tag clouds or other types of search, is what makes them of value to the enterprise. Ok enough of an introduction here is the speech in full. Remember it was given way back yonder in 2000. :)

The sum total of humankind's knowledge doubled from 1750-1900. It doubled again from 1900-1950. Again from 1960-1965. It has been estimated that the sum total of humankind's knowledge has doubled at least once every 5 years since then. It has been further projected that by the year 2020, knowledge will double every 73 days. This statement was made by Patricia Breinik in her book Student Learning in the Information Age. As the information available on the internet increases exponentially over time, so will the value of search and particularly search engines. Learning how to search for answers is important for all of us, whether we do it as students, workers, parents or just for the love of knowledge. I have literally spent years trying out different search engines and watching them evolve since the inception of the internet as we know it today. Our discussion on search engines will center around the types of search engines as represented by their business models, the one type of search engine that has proved to be a search engine in the purest sense and a look at a successful business model of this search type. But first, what types of search engines are there?

Well, search engines can be classified into two major categories or types of searches. These categories are the paid search rank and the relevant search rank. The paid search rank is where advertisers pay to be seen at the top of a search query. The company that pays the most for the position will receive the number one ranking and so will have a better chance of having someone click through to their site. This serves the purpose of generating cash flow for the company who owns the search engine as well as for those that pay to have their websites ranked high on a search query result as they will receive greater traffic than they could generate on their own. Many portals or (definition here) started off with this business model and eventually became hubs of links know as portals such as yahoo, excite and alta vista all of whom are now portals giving users a massive starting point of links based on various subjects.

The second type of search is the relevance search. This is a search that bases its results on the relevancy of a pages answer to your query. (quote from cnet news.com here) With this method, pages are ranked according to the relevance to words, content and links to other sites. A voting method is employed in which a web page or site receive points when other sites link to it. Those other pages are voting that the page they are linking to has value. An example of the difference between the two types of engines is the query of "Canada". Entered at a paid search site you would receive the first five pages filled with travel agency links promising the best airfare prices to Canada. Enter your search term of Canada at a site that judges relevance and the first return would be Canada.gov, the official site of Canada the country. Can you see the difference?

Well, what types of engines are out there that employ relevance as a search model? One of the most comprehensive relevant search engines is Google.com. They invented and patented their own brand of relevant searching called PageRank which was described earlier. Google conducts 150 million searches a day for users. The google search engine is an index of over 2 billion web pages and is updated constantly. The search engine that indexes the next larges number only indexes 7% of the number of pages that Google does. The engine is available in 35 languages and is so good at language translation that if there is a relevant site to your search in another language or country, you can choose to have the page translated for you on the fly. Google is profitable, debt free and privately held, a testament to its business model of paid keyword adds that are viewable down the side of the page but do not count towards page ranking.

As we have seen, the value of search cannot be underrated. Considering the types of search engines and the business models behind them can help us make the right choice. The value of search comes down to relevance. Do you want to be sold a product or are you after the results that best match your search query? Search is the filtering of knowledge...and knowledge is power. Ask yourself this question. When the sum total of humankinds knowledge doubles every 73 days, will you have the power to harness it?

Being unfamiliar with REST'ful applications(or so i thought till i read that this is just a new name for something we've been doing all along..like ajax), I had decided that I would probably be writing some kind of wrapping class for the twitter calls. Enter the yedda api's that are freely available here. I had discovered them some months back and had simply moved them to a todo folder. Little did i realize that when it came down to it I only need to write a few lines of code to use the api to get a simple portlet to post to my twitter timeline.

I plan on using something like the accordian slide from mootools demo'ed here to switch between submitting tweets, viewing timelines, etc. For now though all i have to offer is a cheesy textbox that submits to a timeline. I jinged it below ( i think you can pop it out to a full window if its tiny). Basically all i did was take the yedda dll and start making references. For the demo below all it took was 2 lines of code so you can concentrate on what you want to do with the portlet and how you want it to look. You can also make it wsrp or jsr compliant or create a web service out of it easily. All of this thanks to yedda which is by the way a pretty cool site.

Happy twittering and remember, not every company has an sms gateway lying around so a twitter interface just might do the trick if they want to start going mobile.

Cheesy Twitter portlet demo:
You can view the full resolution version at screencast.com.

First, the more that I work with the 6.5 version of Aqualogic User Interaction (now called Webcenter Interaction) the more I see opportunities for social applications within the Enterprise in general. The Aqualogic portal has always been community centric and social to some extent but as more and more everyday users delve into the social computing world outside of work we are seeing more socially based solutions thought up for common use cases. This is the result of users experiencing new constructs to solve problems. For example, if someone's main form of communication is email and they are familiar with little else they are more inclined to think of an email-ly solution to a use case around communication like support@blah.com. For those who wander into the social computing world outside of work they gain new constructs to solve issues using contructs like sms, rss, activity feeds, etc.

One such use case comes from a governent agency I had the pleasure of visiting in western Australia earlier this year. I was onsite for a two week workshop for estimating level of effort for migrating onto the ALUI portal platform. Part of the workshop involved "gotta haves" while part of the workshop involved "nice to haves".

One of the use cases that they brought up was the need to communicate to non-employees. This agency runs a very large apprentice program so that students who have recently graduated from high school but who will not be attending college can learn a skill or trade. This large group of apprentices is placed through the agency but does not work directly for the agency. They do however have to be communicated to both en masse as well as individually. Many of the locations that the apprentices work at do not have an internet connection or computer at all so there is no way for these users to log in and check email or to log into a portal. We kicked around the idea of using an sms gateway to send activity feeds or one off messages to the group but as a quick hit i mentioned twitter.

One person in the group was semi-familiar with Twitter and had a construct for understanding the concepts. I did a quick POC to introduce the group to twitter. I had them all sms "follow blahblah" to 40404 which registered their phone for the twitter feed. From then on you can post to blahblah timeline and everyone who chooses to follow receives the feed via sms. The simple steps would be to send out a communication (whatever their current method was) that asked all apprentices (vast majority having cell phones with sms capability) to follow the corporate username on Twitter. A portlet would be written that posts timeline entries to the Twitter account and uses Twitter's sms gateway to handle the sms blast to the apprentice group.

As I pretty up the portlet I'll post other use cases that I've seen that lend themselves to some sort of Twitter-like solution. Certainly just in the realm of Enterprise 2.0 facebook-y apps having a way to post to and read from your twitter timeline would be useful. The next post will deal with the portlet itself.

Fortunately Pages 1.5 showed up on the edelivery.oracle.com site today and i was able to download and install without issue. Pages 1.5 uses configuration manager 2.0 to manage its settings so all is right with this DEV world now.