Use Cases & Requirements/Map Application UC

From Media Fragments Working Group Wiki
Jump to: navigation, search


Why it is important to define and identify fragments of an image? Why one should bother to identify regions of an image? Why a server would only deliver chunks of an image?

Map UC

Map applications usually have to deal with large images. The user of the map application can zoom in and out, scrolling to left or right, etc. Furthermore, the user can search for particular regions/locations. In this context, the identification of media fragments is useful for both annotation and delivery purposes, and when sharing the image fragment.

Web Map Services

Web mapping is the process of designing, implementing, generating and delivering maps on the Web. The Web Map Service (WMS) protocol offers a standardized method to access maps on other servers. WMS servers can collect these different sources, reproject the map layers, if necessary, and send them back as a combined image containing all requested map layers. One server may offer a topographic base map, while other servers may offer thematic layers.

WMS server are specialized web mapping servers implemented as a CGI application, Java Servlet or other web application server. They either work as a standalone web server or in collaboration with existing web servers or web application servers (the general case). WMS Servers can generate maps on request, using parameters, such as map layer order, styling/symbolization, map extent, data format, projection, etc. The OGC Consortium defined the WMS standard to define the map requests and return data formats. Typical image formats for the map result are PNG, JPEG, GIF or SVG. An open source WMS Server is the UMN Mapserver. Commercial alternatives exist from most commercial GIS vendors, such as ESRI ArcIMS, Intergraph Geomedia WebMap and others.

A comparison of the three most important web map services (Google, Yahoo! and Microsoft Live maps) is available at

The Open Geospatial Consortium (OGC), previously known as the Open GIS Consortium, is an international voluntary consensus standards organization aiming at encouraging development and implementation of standards for geospatial content and services, GIS data processing and exchange. Most of the OGC standards are based on a generalized architecture captured in a set of documents collectively called the Abstract Specification, which describes a basic data model for geographic features to be represented. There are currently 28 standards in the OGC standards including WMS, GML, KML, etc.

Google Maps,+avenue+du+G%C3%A9n%C3%A9ral+de+Gaulle,+06212+Mandelieu,+France&sll=37.0625,-95.677068&sspn=32.38984,79.101563&ie=UTF8&ll=43.532932,6.947865&spn=0.014467,0.038624&z=15&iwloc=cent

Yahoo! Maps

Microsoft Live Maps

  • Web:
  • The Live Map URI generated for the address 605, avenue du Général de Gaulle, 06212 Mandelieu, France is:


Multi-resolution Images UC

A specialization of this use case would be to identify and retrieve zones at different resolutions such as individual tiles or set of tiles in a multi-resolution image. The whole image industry is willing to deliver high resolution images ... by parts, but only if it is hard on the client side to "reconstruct" the original and complete image.

This concept is popularized by the Zoomify product that you can find in a lot of applications including GoogleMaps and siblings. The idea is that you have a fix size window on the screen to display a large image, so you will need to visualize it by parts, and have some zoom in / zoom out functionalities. Each parts could be associated with a URI defined by the WG. Zoomify has developed its own format and sell the source code of a flash application on top of this model ... but people told me that having a URI syntax for identifying in a generic way these parts would be better. A number of open source softwares doing similar things have been released recently, and we will not have problems to have implementations.