The user interface for the search engine

The user will assume that the system of "their" if it is to a large extent can control the results of its work, by introducing the control actions and seek feedback. In the case of search engines impact the user - this is a query and a set of additional parameters. We can state that can effectively control the search results in general are inversely related to the volume of the index system.

For example, none of the known systems, the author is not in a position today to rank the documents found on the length, as did the former system Alltheweb, at a time when the index does not exceed one billion documents. The system of "Yandex", volume index is at present - 4.55 billion documents, allows us to rank relevant documents by date, which is not able to do Google, whose index is orders of magnitude greater. Namely, the latter possibility seems to the author of fundamental importance for the analytical work - at a constant monitoring of publications on the same request. For the user, essentially the presence of intuitive information retrieval language, easy navigation of search results, rankings, results are automatically generated understandable summaries of documents (snippets).

Interface Modes "advanced search" of most search engines today, on the one hand, overwhelmed, and the other - do not allow to fully express all the information needs. Very often the user is literally drowning in the results of the primary search, but, going into "enhanced" by filling out all applicable fields, comes to the lack of results (despite the fact that the main problem of the Web as an informational medium - this is redundant information). In this regard, recently received a spread of more flexible visual interfaces clarification requests, usually implemented through "kvaziintellektualnoy" grouping (clustering) of the results of the primary search. There was a lot of approaches have in common that - trying to present search results and the corresponding clusters in the user-friendly form.

For example, Russia's metasearch system Nigma.ru for the word "Internet" is clusters (grouped documents), corresponding to certain words automatically (tag) "online store", "services", "catalog", "delivery", "link" etc. ., indicating that the commercialization of the Russian segment of the Internet. Nigma.ru also uses clustering results are displayed in the form of "cloud" of tags. On each tag can "click", resulting in a subordinate tags. Search results corresponding to the selected tags are displayed in the bottom of the page in the traditional form of list.

Another Russian search engine Qintura (http://search.quintura.ru/) has an "intelligent" interface, which provides automatic hints to your query, helps to dynamically manage the search process. At the request of one word Quintura makes possible sentences and phrases, which, if necessary expand the primary search.

Developed graphical user interface system has Grokker (grokker.com), results which are grouped in a pie chart. The system can search databases of Yahoo!, Wikipedia, Amazon Books. An interesting feature of the system are such tools refine the query as a time management publications (with a "slider", as well as a well-known news aggregator NewsIsFree), as well as the choice of domains or sources.

One of the most dynamic news resources of the Internet today can be considered alive and logs (blogs). The company TouchGraph, in particular, to implement an interface to visualize sotsiosetey on Livejournal. TouchGraph LiveJournal Browser. However, the most popular tool of TouchGraph Google Browser (http://www.touchgraph.com/TGGoogleBrowser.html), which is a Java-applet for visualizing thematic similarity of Web sites. This is a very useful tool when searching for sites related to the overall theme of the original. The interface TouchGraph Google Browser shows all the sites linked to the original references given site, and the user can specify the depth of relationships and show how the different sites.

System for visual search focused on the psychological aspects of human perception, focusing on techniques that people use during the search. That is why visual search engines have all the chance to excel in the information market giants such web search like Google and Yahoo!, using a database of the latter. On the other hand, may be more obvious course is to move the leading network IPS in visual search? After all, users visually organize the search results look much more attractive and easier to understand than lists of hyperlinks and snippets generated by traditional search engines.

Today, the bulk of information resources on the Internet, not in the Web sites and in the so-called peer to peer, file-sharing networks. In practice, peer networks consist of workstations, each of which interacts only with a subset of nodes in the network (due to limited resources). To implement the protocol used P2P client programs that enable both individual workstations and the entire peer network as a whole.

Quite often, peer networks are complemented by dedicated servers. It is often these servers can address search requests, as this is the problem for peer to peer networks can not be solved. The question of an effective search in such networks by means of the networks themselves is still open (as most of them does not imply a rigid centralization, but rather, they are by definition decentralized), but there are special search sites in the web space to help solve this problem.

When searching for a peer to peer networks, subject of a search overshadowed the main problem is - quickly and efficiently find the most relevant response to a request transmitted from the workstation across the network. In particular, the urgent problem - reducing the network traffic generated by the query (for example, sending a query to multiple workstations), and at the same time to obtain the best performance issued documents, ie obtaining quality results.

Acceptable quality search peer networks today provide a specialized, centrally filled, search websites, of course, running over HTTP. For example, for eMule file-sharing network in such a search engine is the site Figator.com, and to network Bittorrent - site isoHunt.com.

As for file-sharing networks, for those servers to be particularly relevant and critical issues are the quality and reliability of content, fabrication and distribution of false file resources, viruses, "Trojan horses", the possibility of falsification of ID workstations.

© 2010 - 2019 D@nVitLabs