O’zaro ta’sirli interfeysni loyihalashda hisobga olinishi kerak bo‘lgan operatorning asosiy xususiyatlari

2.6 Techniques for the Adaptation Phases
In automatic adaptation we can identify three main phases: device identification, interaction resources identification, adaptation.
Device identification can be performed either server-side or client-side. In the server-side case, usually the user agent detection in the HTTP Protocol is carried out. It normally indicates: ([system and browser information]) [platform] ([platform details]) [extensions], for example: Mozilla/5.0 (iPad; U; CPU OS 3_2_1 like Mac OS X; en-us) AppleWebKit/531.21.10 (KHTML, like Gecko) Mobile/7B405. In the client-side case, some level of identification of the main features of the current device can be performed through the markup (for example, the srcset attribute is able to indicate which version of an image to use depending on the main features of the device); or by using the stylesheets, which are associated to different devices by using the media queries; or by using certain scripts (e.g. jQueryMobile provides support for this purpose).
Interaction resources identification is applied when it is necessary to have more detailed information on the currently available interaction resources. The environment should then access them by using a Device Description Repository (DDRs). One format for DDRs is given by the UAProf (User Agent Profile), which describes the capabilities of a mobile handset, including screen size and multimedia capabilities. Mobile handsets send a header (usually "x-wap-profile“) within a http request with the URL to its UAProf. Its production for a device is voluntary. It is based on the Composite Capability/Preference Profiles Specification (CC/PP) created by the World Wide Web Consortium. Another format for DDRs is given by WURFL, which is an XML configuration file that can be stored locally and contains information about capabilities and features for a variety of mobile devices. This information is derived from different sources: UAProf, when available; public documentation; developer reports; actual testing. It has a hierarchical, extensible, structure. It started as an open source project http://wurfl.sourceforge.net/, and now ScientiaMobile, founded by the WURFL team, offers commercial support for these APIs, also as a cloud service. Other commercial tools in this area are Device Atlas (http://deviceatlas.com), and DetectRight (http://www.detectright.com).
In general, the device properties can be classified as either static, which cannot change during application execution, such as operating system, RAM size, available storage, display size, input devices, markup support, CSS support, image format support, script support, etc.; or as dynamic, such as device tilting, network technology in use, quality of connection, battery level, location, orientation, acceleration, light, noise, etc. Media queries are able to detect a limited set of media features: width, height, device-width, device-height, orientation, aspect-ratio, device-aspect-ratio, color, color-index, monochrome, resolution.
The third phase is adaptation. There can be various approaches to automatic re-authoring:

• 
  Scaling: just linearly scaling according to the interaction resources of the available device, such as Safari on iPhone does when loading a Web page developed for desktop systems;
  
• 
  Transducing: preserves initial structure and translates elements into other formats, and compresses and converts images to match device characteristics;
  
• 
  Transforming: goes further to modify both contents and structures originally designed for desktop systems to make them more suitable for display on small screens.
  

The problem of performing an automatic adaptation from a desktop to a mobile version able to change the user interface structure can be addressed by first calculating the costs in terms of screen space required by the various elements: i.e. the vertical and horizontal space required by a text, image dimensions, interline value, interactor type, etc. Next, calculating the space required by the user interface in the target device should also consider how much tolerance in scrolling should be allowed, how much additional space should be available for tables, and similar aspects. If the result is higher than the sustainable cost for the target device then the adaptation of the user interface elements should be considered (e.g. using smaller images and replacing interactive elements with equivalent ones that take less space). If the resulting overall cost is still excessive for the target device screen then splitting the user interface into multiple presentations should be considered. In order to decide how splitting into multiple presentations should be performed the user interface can be considered as a set of groups of elements, which cannot be split internally. Thus, the decision is how to distribute such groups in order to obtain presentations sustainable by the target device. Splitting can be implemented either creating separate mobile presentations or by showing the relevant elements dynamically. This adaptation process can be customized according to certain parameters and rules, such as how much scrolling should be allowed in the target device or what policy to follow in distributing the groups of elements in the target device. In this adaptation process sometimes tables are critical elements because when they are shown on the small screen device they are too large. Some techniques have been proposed to handle such issues, for example Tajima and Ohnishi (2008) introduce dynamically scripts that allow some columns and/or rows to be collapsed interactively in order to enable users to relate better the elements of interest in the table.
Another interesting adaptation technique is page summarization, whose purpose is the automatic reduction of content in order to make it more suitable for small screens. There are two types of approach to this issue. The Abstraction-based approach uses sentence manipulation techniques like reduction, compression and reformulation. The Extraction-based approach assigns scores to sentences in order to select those which better represent the whole text; it can be feature-based (e.g. term frequency, sentence position, attributes, etc.), or make use of machine learning or graph based techniques.
An example of summarization is that supported by PowerBrowser (Buyukkokten et al., 2002). The basic idea was that the importance of a keyword depends on the frequency with which it occurs in a text and in a larger collection. A word within a given text is considered most important if it occurs frequently within the text, but infrequently in the larger collection. The significance factor of a sentence is derived from an analysis of its constituent words. The sentences in which the greatest number of frequently occurring distinct words are found in closest proximity are probably important. People interested in such techniques can use MEAD, a public multi-document summarization system, which provides more flexible support in this area (see http://www.summarization.com/mead/).
Crowd-sourcing techniques are based on the idea of allocating some tasks to perform through an open call. These techniques are acquiring increasing importance and can be applied to adaptation as well. For example, Nebeling and Norrie have applied them to adaptation of Web pages. The goal is to support developers in specifying Web interfaces that can adapt to the range and increased diversity of devices. For this purpose they have introduced a tool that augments Web pages to allow users to customize the layout of Web pages for specific devices. Devices are classified in terms of window size, screen resolution, and orientation. It is then possible to share adaptations so that others with the same device and with similar preferences can directly benefit. The same group ( Nebeling et al, 2013) has developed a tool, W3Touch, whose purpose is to support adaptation for touch according to metrics. The tool produces analytics of the user interaction in order to help designers detect and locate potential design problems for mobile touch devices. For this purpose two metrics are considered: Missed links ratio, which keeps track of how often touches miss an intended target; and Zoom level, which considers how much users on average need to zoom into different components of the Web interface.
Another important aspect to consider is how to evaluate adaptation. For this purpose in Manca et al. (2013) a set of relevant criteria are indicated:

• 
  User’s awareness of adaptation: to what extent the user is able to realise that a change in the UI is caused by adaptation;
  
• 
  Appropriateness of adaptation: whether the system selects a good/appropriate adaptation strategy;
  
• 
  Transition of adaptation: to what extent the adaptation process allows users to realise what is happening during adaptation;
  
• 
  Continuity of adaptation: to what extent it is easy to continue the interaction after adaptation;
  
• 
  Impact of adaptation in decreasing interaction complexity: whether the interaction complexity of the system decreases;
  
• 
  Impact of adaptation in increasing user satisfaction: to what extent adaptation increases the user’s satisfaction.