A Report on the Activities of the Sahana Japan Team after the Great East Japan Earthquake
【Introduction】
In the aftermath of the Great East Japan Earthquake, which occurred on March 11th, 2011, volunteers involved in the open source community formed the Sahana Japan Team to engage in activities to utilize Sahana in relief efforts to assist the disaster victims.
By reviewing the activities carried out by the Sahana Japan Team during this period, this document describes how these activities were conducted and what kind of issues were encountered, with the aim of providing insight for individuals involved in similar relief activities in the future.
【Contents】
- Launching of the Sahana Japan Team
- The Activities after the Launching of the Sahana Japan Team
- The Development to Utilize in Japanese Language Environment
- Releasing the Japanese Version of Sahana and the Usage of Sahana in the Devastated Areas
- Translation Process
- Public Relations
- The Assesment of the Activities
- Accomplishments
- Improvement and Reflection
Launching of the Sahana Japan Team
Sahana was first introduced in Japan at the “Open Source Conference 2010 Kobe (OSC)” and the “International Symposium on Disaster Reduction.” These events led the members of “Hyogon Tech,” a volunteer group in Kobe, to begin work to localize Sahana into Japanese and to promote its use in Japan.
“Hyogon Tech” is a group which aims to support civic activities from the aspect of information and communications technology (ICT). Hyogon Tech specifically focuses on supporting the usage open source software as a means to facilitate the activities of NPOs and private citizens.
After learning of how open source software was utilized to support relief efforts for the Indian Ocean earthquake, one of Hyogon Tech's members invited Mr. Ketut Ibnu Sutawijaya, a member of “COMBINE Resource Institution,” an Indonesian NGO, to give lectures at the 2 aforementioned symposiums on how Sahana was utilized in Indonesia (held in association with AMARC Japan).
This meeting acted as the catalyst for the formation of the “Localizing Sahana into Japanese Project,” which was a predecessor of the “Sahana Japan Team (SJT),” and efforts to translate Sahana into Japanese commenced.
At the time, although two versions of Sahana existed (a PHP version called Agasti and a Python version called Eden), it was decided that only Eden would be localized into Japanese. The reason for this is that Sahana Eden utilizes the web2py framework to provide functionality that switches the display language based on the language settings specified by the user's browser, allowing for usage in various languages including Japanese without the need to perform any additional settings.
Additionally, Hyogon Tech periodically organized “Tech Cafe” events to announce updates on the Japanese localization process and to promote the usage of Sahana, etc.
In the aftermath of the Great East Japan Earthquake, which occurred on March 11th, 2011, volunteers in the open source community began activities to implement Sahana to aid in the relief efforts, eventually leading to the formation of the Sahana Japan Team as an open source development community.
On the day of March 11th, 2011, members had been gathered to conduct a meeting in preparation of OSC 2011 Kobe. Among the participants were members of the “Localizing Sahana into Japanese” project, who in response to the earthquake, decided to begin efforts to implement Sahana to aid the disaster victims.
Additionally, Mr. Chamindra de Silva, the CTO of the Sahana Software Foundation (SSF), which oversees the main development for the Sahana platform, contacted "Hyogon Tech" to offer support. Furthermore, on the day following the earthquake, a multitude of volunteers contacted "Hyogon Tech" via mailing list to offer support in such things as aiding in entering data to the server, testing, and additional localization work.
Activities after the Launching Sahana Japan Team
Development to Facilitate Usage in a Japanese Language Environment
Initially, team members had thought that in order to use Sahana, the only thing needed was translation into Japanese. However, the version of Sahana Eden available at that time contained bugs that prevented the application from being used properly in a non-English environment, thus necessitating SJT to focus on fixing such bugs in the development version provided by SSF.
Selecting the Development Version Used for Deployment
For using Sahana Eden in Japan, the Team first attempted to use the newest and stable version (0.5.4) provided by SSF, but later found out that its usage in Japan would prove to be too arduous because of issues encountered during initial setup preventing proper configuration in a non-English environment.
For this reason, the team opted to perform setup of the testing environment for japan.sahanafoundation.org by installing a development version obtained from the development repository.
Repeated Coding and Fixing
The development version contained some bugs. The SJT began their efforts by focusing on fixing these bugs.
While the SJT was performing such debugging and additional bug fixing, the core SSF team continued to add new revisions to the same development repository, resulting in a vicious cycle of fixes and revisions to maintain compatibility. However, as the team considered the features added and improvements made by core SSF team to be useful, the team continued tracking the development trunk.
IBM Japan’s Participation in the Project and the Launching of the Japan Branch
Although the IBM Japan Team later aided the Team in documentation and the testing process, due to the worsening of the aforementioned lack of coordination with main development, many issues occurred, such as additional revisions being made to code that had already been tested and changes made to the interface which affected existing documentation. Such issues lead to dire circumstances for the Team.
For this reason, it was decided that the team would form a separate Japan branch of Sahana (focused on deployment and usage in Japan) from the development trunk provided by SSF, and that efforts would focus on bug fixing for the new branch.
For creating a new branch, the team opted to select the newest development version of the software available at the time, with the intention of maintaining the progress gained from their efforts leading up to that point. This decision, however, necessitated the team to make revisions to unstable and incomplete code, resulting in a little over 2 months of additional work to make the system robust enough for practical use (pushing back the release date).
Releasing Version 0.1 and the Usage of Sahana in the Disaster Affected Areas
Facing such circumstances, based on a proposal by IBM Japan, the team decided to focus on development and testing of high priority functions, culminating in the release of Sahana jp0.1, a bare bones release providing the minimum necessary features.
The bare minimum features provided by the first release include the “Shelter” module used to verify the conditions of evacuation shelters, the “Req” module used by victims to request assistance, the “Inv” module used to manage supplies that have been requested, and the “Map” module used to map the aforementioned information.
Subsequently, IBM Japan provided further assistance by making further improvements to the program, entering basic information into the database, performing setup of the tablet computers distributed to the evacuation shelters, and making user manuals for the application. Through such efforts Sahana was deployed in Iwate and Yamagata Prefecture, and the system was used for managing the victims at the evacuation shelters, requesting goods and health services, and for surveying the conditions of the evacuation shelters
（#for more details on the usage situation in Iwate and Yamagata Prefecture, please refer to “A Report on Field Research Conducted on the Usage of Sahana in the Tohoku Region).
Translation Process
Pootle, which is an online language catalog translation system, was used to perform translation from English to Japanese. Additionally, documentation provided at the SSF website was translated and published on the Sahana Japan Team’s website.
- After the launching of the “Localizing Sahana into Japanese Project," translation to Japanese began in March 2010.
- The usage of Pootle made it easier to merge multiple versions of the same document edited by multiple users.
- After the occurrence of the East Japan Earthquake, IBM Japan provided assistance to help complete the translation process into Japanese.
Public Relations
To gain widespread support, participation, and cooperation for the Sahana Japan Team, the project members actively attended IT related events, such as OSC, in order to introduce Sahana to those in attendance.
- The Team introduced Sahana at OSC 2011 Kobe held on April 16th, 2011.
- The Team introduced Sahana at OSC 2011 Sendai held on May 21th, 2011 and attended Hack For Japan held in Sendai on the same day.
- The team continues to disseminate information through documentation, reports, and calls for support, via the SJT website.
Assessment of the Activities
Accomplishments
As an accomplishment of the Sahana Japan Team, the team was able to make improvements to internationally developed software that had proven to be difficult to use in a non-English environment, and made it suitable for usage in a Japanese language environment.
With assistance from IBM Japan, Sahana was deployed and used at evacuation shelters by several local governments such as Iwate and Yamagata Prefecture to manage the distribution of supplies as well as information on evacuees, enabling the Sahana Japan Team to contribute to the assistance of the earthquake victims.
In addition to fixing bugs related to multilingualization, many of the bugs fixed by the team were reported to the Sahana Software Foundation, and were incorporated into the main development trunk, contributing to the worldwide development of Sahana.
In addition, the activities of introducing Sahana at OSC and other events could spread the existence of “Sahana,” as open source software for the evacuees of disaster especially among IT concerns.In addition, by introducing Sahana at events, such as OSC, the team was able to raise awareness on Sahana's role as an "open source software platform used to facilitate disaster relief efforts," especially within the IT community.
Issues Encountered and Lessons Learned
Using Sahana in a Japanese Language Environment
Since Sahana is well-renowned for its long record of past deployments for overseas disasters, team members were under the impression that the software could likewise be deployed immediately after the disaster in Japan. However, many problems were encountered related to the conversion of multi-byte characters, and in order to make Sahana usable in a Japanese language environment, the team was required to perform extensive revisions to the software itself, in addition to the anticipated translation work.
Due to the large amount of time spent on such revisions, the team was unable to deploy the software in the disaster affected areas, in the immediate aftermath of the earthquake, at the time when it was needed most.
Additional issues were faced when considering how to translate such attributes as "firstname" and "lastname" into Japanese. If translated literally, the order in which the name is displayed would run counter to the manner in which Japanese names are traditionally listed (i.e., last name first, first name last). Furthermore, if translated in the reverse order, the issue then arises of how to handle the names of non-Japanese residents living in Japan. As shown here, it is important to pay attention not only to differences in languages but to cultural differences as well.
In the future, when adopting Sahana for use in disaster relief efforts in various countries, it is evident that thorough testing will need to be performed in advance, under the usage environment specific to that region.
Additionally, regarding the translation process, limitations were encountered with Pootle that resulted in questionable translations, as certain English words possessed different meanings depending on the context they appeared in.
Communication Problem with Sahana Software Foundation (SSF)
The SSF core team wanted to support SJT better. However, due to the language barrier between the SSF and the SJT, poor communication persisted.
For the SJT, the immediate priority was to fix any bugs to ensure stable operation so that the system could be prepared for deployment in disaster relief efforts. However, the difficulty to coordinate simultaneous development between SJT and the core SSF team delayed efforts to deploy Sahana for practical usage.
It was observed that main development at the SSF has been advancing at a fast pace, and that there appears to be a propensity to add new features and to focus on code refactoring, rather than to put emphasis on maintaining compatibility.
This brought the completely breaking compatibility between the branched Japanese version and the main development trunk in a matter of months, making it difficult to merge any code developed or revised in Japan to the main development trunk.
The same issue has been identified for local branches outside of Japan (such as in Taiwan, etc.), making it difficult to share code and to contribute to the progress of development between countries.
Lack of Management
Since the activities conducted involved such members as development engineers and volunteers, etc., who had been assembled on very short notice, in the ensuing period after the earthquake, the team was found to lack proper management capacity and even calm judgment at times. This was especially evident when the team was required to collaborate with private sector counterparts, and this proved to be a detriment when trying to coordinate work procedures and work progress between those involved. Furthermore, there were also cases where the team was not able to properly handle feedback and offers of assistance received from volunteers.
It is important to note, however, that there were some benefits of adopting a bottom-up style of management based on the self-motivation of individual members, and that a certain degree of confusion is an inevitable consequence for such projects conducted in emergency situations that involve volunteer members. With that said, when conducting similar activities in the future, it may be necessary to improve management methods by implementing visualized task management, visualized allocation and sharing of tasks, and by organizing a follow-up system to handle participants who join the project when it is already underway (for organizing documents and facilitating lecturing methods).
Others
Although the team was able to raise awareness of Sahana within IT/OSS circles, the team was unable to sufficiently raise awareness among the organizations and groups that would actually be required to make practical use of the system. Additionally, the team was unable to receive enough information about the needs and the conditions in the disaster affected areas.
As an additional note, the psychological stress arising from such tasks to indirectly support the disaster affected areas was immense. Organizations involved in similar activities are well-advised to take into account the mental state of individual members involved in such activities.
【Acknowledgments】
After the occurrence of the earthquake, the Sahana Japan Team has received vital support from numerous volunteers, companies, and organizations. The Sahana Japan Team would like to express its gratitude for all cooperation provided and hopes that its efforts will contribute to future disaster relief activities.
2012 Sahana Japan Team
This work is licensed under the Creative Commons Attribution-ShareAlike 3.0 Unported License
