Search Engine as the most valuable enterprise application component.

I have been very quiet for very long time. It was not because I have nothing interesting to report. On the contrary, it was a roller coaster ride of application development. This year can be labeled something like “Rich Internet Applications meet Enterprise Database by the way of full text search”.
I have been working on the projects to help enterprise subject matter experts do their work. It was very addictive and ever changing process that required reaching certain level of results followed by a cool review to make it believable to a “sane” developer.
We are slowly getting used to the fact that of-the-shelf and free software can find related data out of million records within a second. The fact that the found data can be ordered and presented in a form of the definite answer recognizable by user in a single glance is nothing short of magic.
Enterprises store vast amount of information comparable in size to the public accessible web – some of the clients few years ago were actually bragging how much more data they have over Google. However, most of the workers cannot access it due to data size and “system walls”. As a result, subject matter experts are mainly being used in their historical capacity – people who memorize the facts and working within those constrains. Full text search, combined with some metrics on the quality of the data, challenges both what we “know” and how we perceive our “knowledge”.
The way the knowledge is processed is about change in the next few years, as the companies really have to open up these databases for use if they want to maintain competitive edge. It will lead to a revolution in the enterprise application development. Let us see few side effects of that process so we can recognize the change.
The use of full text search will have a major impact on how the applications interact with the data. Instead of actively looking for information by digging for information with tons of drill downs and clicks, you just type what you want or highlight what matters to you.
The system uses those queues to search, rate and present information back to you with highlighted things that match your request. Moreover, with proper training systems can also identify what might be the reason why it is not applicable in the context.
Properly designed “data driven” application adapts to the things you usually search for as well as data it searches on. It can let user to add as many qualifications or “disqualifications” to the search as you seem appropriate without usual constraints of the structured data. Once a certain level of experience is established, the system easily outperforms the subject matter experts in the speed and quality of data retrieval and statistical qualifications. The main role of the human shifts to use of their analytical and pattern recognition skills.
Flex is really ideal environment for this type of applications, as you really want not only “refresh-free” experience but really “seamless” one. The ability of creating seamless environment is crucial, as the natural system implies user experience that adapts to the process AND context. As a result, visualization and transparency of the context and state are as important as quality of the search engine.
“Seamless” also implies natural flow of things. The push technology that Adobe was promoting for so long is finally coming to enterprise applications.
I will cover in next parts few components that are coming from all that development. Good news is that SEO2 are within the reach and in the eyesight of the new crop of applications.
Anatole Tartakovsky