Search engines are the gateways into cyberspace. These are robots that scour the Web to index the content. They form the largest single area of web activity (80%), followed by information (76%), weather (65%), bill payments (56%) and price comparisons (50%) (2013).
Since the advent of computers, the structuring of information has always sought to meet the challenges of the three Vs : Value, Speed and Variety:
The first generation
Appears in the early 1990s : Archie in 1990, Gopher in 1991, WebCrawler and Lycos in 1994, Excite, Yahoo and Altavista Netscape in 1995 and in 1996. The first search engines index content corresponding to the words contained in a query.
The second generation
Begins in 1998 with Google PageRank that provides an index of ranked notoriety. Google then experienced tremendous growth because its approach was linked to an advertising strategy. Such research does not necessarily refer to the main reference on the subject, but rather to the most-often referenced, via algorithm.
The third generation of capability is being built
- it will operate streaming through browsers that will analyze the data stream. Thus, people will not try to connect to a particular site but instead tap into a particular flow of knowledge ;
- it will use geo-location and traceability to customize more and more research, and will favor (short-term) development of niches and (long-term) the development of local economies ;
- it will also operate effectively with mobile devices (chapter 4, no 11) ;
- it will address not only numerical data, but also sound, voice, images, video, surveillance cameras, card payments, archives, social networks, shopping sites, etc.
- it will seek to process and translate requests in natural language (chapter 5, no 7).
There are two common sorts of Predictive Intelligence :
- The first uses data and text mining to identify, by reading statistical data and keywords, patterns that are heavy or light (chapter 4, no 4).
- The second uses charts and diagrams to establish pattern recognition ; it is the form of predictive intelligence that uses the intelligence of the eye (chapter 9, no 5).
This movement that is building this important predictive approach is fed by several streams : Predictive Analytics, new neuroscience research (see BRAIN projects in Chapter 3) or the development of a Predictive Model Markup Language.
An example of predictive intelligence which became a tragic error :
An Iranian Airlines Airbus was shot down by the cruiser Vincennes on the 3 July 1988. This cruiser had a computerized detection-and-pursuit system (Aegis). Its command center was in a room without window-studded screens on which were displayed airplanes and ships within two hundred kilometers away. Any activity or movement appears in the form of a symbol (see below).
It remains surprising that such a costly system (its development cost over a $ billion), capable of killing people at a distance, was so primitive in terms of the use of icons.
A second example
In 1911, we asked the watchmen of the early century to describe a lounge of the next century. Because the phone and the radio then emerged, the watchers thought only in terms of audio.
In the engraving below, the living room is filled with speakers connected to a multitude of wires, each offering a different source of information from a different city. As for the image (so important today), it is reduced to a small dancing couple (center):