Position Paper for W3C Workshop on The Future of Off-line Web Applications

5 November 2011 Redwood City, CA, USA

Offline Web Applications for Disaster Relief

Satoru Takagi , KDDI

Introduction

We strongly realized the need of the offline web application by the great earthquake disaster that occurred in Japan this year.

We understood that we could prevent the damage of servers and backbone transmission channels to some extent by redundant deployment. However, this can not be applied to the user terminal side. Keeping a terminal connected to the Internet is very difficult at the time of disaster. Browsers disconnected from servers are totally useless for a user. It was a user not a server who is damaged most by a disaster. We wish WWW works in such situation effectively, since one of the greatest features of the WWW is gathering and integrating diversity of information and providing a user with a single comprehensible information.

Usecase Requirements

At the time of disasters, the form of the following information exchange is effective.

• The handiness that ordinary people in the suffering place can build information without the support of the servers
• The handiness that ordinary people in the suffering place can broadcast information without the support of expert engineers
• Compliance to the information exchange with the all sorts of networks such as mesh-shaped ad hoc network like P2P, besides the more primeval network such as UUCP using SD memory etc.
• Securing reusability of data for the use beyond the original purpose of the service. This will be indispensable to utilize data in an unexpected situation to the maximum.

Needless to say, map is particularly important among all the media in a disaster area.

Architecture Requirements

The following architecture requirements will be derived by the requirements mentioned above.

• Information should be serializable, and able to be stored in local storage.
• Information should be understandable and manageable for ordinary people. The realistic candidate would be human understandable web content files stored in hierarchical file system.
• Reusing data effectively by enabling a user to easily gather and integrate diversity of offline information. This is equivalent to the mash-up or hyper-linking by the user-side.

The current dynamic web service cannot support these requirements. The reasons are as follows:

• Interactive communication with server is necessary for the use of current dynamic services.
• The management of current dynamic services requires specialists or experts.
• Serialization of information is complex and difficult for current dynamic services.

Even if dynamic services are limited to REST+AJAX, it is still insufficient. This is because it is difficult other than a person knowing the contents of the service well to find out the data which should serialize. URI of the data which should be serialized is not written clearly as data. It is buried in the logic in JavaScript. Furthermore, depending on the design of the logic, there is even the case that the acquisition of infinite data is necessary to use service offline.

Rather classic static WWW has superior fitness. The reasons are as follows:

• Static Web contents are serializable in hierarchical file system surely.
• Static Web contents can be produced by stand-alone authoring tool such as OpenOffice easily.
• Data which needs downloading can be found out easily by following hyperlink.

Based upon the above facts, we believe additional manners shown below are necessary for Web Application for disaster relief in addition to REST+AJAX.

• Data (web contents) and the logic ( java script programs ) should be separated clearly.
• Also it should be suppressed to bury the connection between data in a logic. All connection should be specified as URI in data. (That is hyperlink)
• The data should be built as the human understandable presentation.(That is web contents)

That is, the WWW data for disaster relief should be built as linked presentation data.

Offline Web Mapping Application

And web mapping application which is required for disaster relief most has many problems mentioned above. The reason is that, current Web mapping services have been created based on Web2.0 and dynamic Web Service architecture. We have the innovative technology that can solve most of these requirements. And we suggested the technical specifications to the W3C in this August as member submission, "Tiling and Layering Module for SVG" (SVGTL).

SVGTL realizes basic functions of web mapping including the mash up without dynamic server and AJAX logic of the browser side. In other words, SVGTL realizes functions necessary for web mapping without programming by using human understandable presentation data for WWW. The core of this architecture to handle very large map which have no choice but to divide into a large number of files and to realize mash-up of diverse information is extended hyperlink.

Importance of Hyperdocuments

We believe that the most important value of the WWW is a hyperdocuments. The reform of the web application except the map should be realized by the innovative version up of the hyper document, too.

References

• Submission: Tiling and Layering Module for SVG 1.2 Tiny Submission
• Documents: Tiling and Layering Module for SVG 1.2 Tiny
• W3C Staff Comment