Wikimedia CH/Grant apply/Dico des Ados offline

From Meta, a Wikimedia project coordination wiki

Infodata[edit]

  • Name of the project: Dico des ados offline
  • Amount requested: CHF 35'000.-
  • Type of grantee: Organization
  • Name of the contact: Stephane Coillet-Matillon
  • Contact: contact(_AT_)kiwix.org

The problem and the context[edit]

What is the problem you're trying to solve?[edit]

Le Dico des Ados is a mediawiki-based dictionary serving a younger, French-speaking audience (ados being teens). It is cross-referenced with the Wiktionary and partly developed in Switzerland.

While most Swiss teenagers nowadays own a smartphone[1], their bandwidth and available data are overwhelmingly dedicated to social media and video consumption[2]. In other words, for a given monthly allowance of data and faced between interacting with their friends or learning something, there is a high likelihood that teenagers would favour the former over the latter. Hence an issue of effective reach for the Dico.

What is your solution to this problem (please explain the context and the solution)?[edit]

Kiwix is the default platform to help bring internet content to people without internet access (still 4 billion people in the world, for whom lack of connectivity may be a matter of lagging infrastructure, or data costs. Kiwix functions much like a browser, reading local and integral copies of material that is normally found on the internet, e.g. Wikipedia, the Wiktionary, or TED talks - freely-licensed educational content in more than 100 languages. The Kiwix software runs on all devices, be they mobile or desktop computers, as well as servers and small, inexpensive hotspots.

In the case at hand we have teenagers, smartphones, teenage-oriented content but no or little available connectivity to navigate said content. This situation sounds very much akin to what Kiwix already is dealing with in the offline world. Hence, we offer a very Kiwix-like solution of bringing the Dico des Ados offline, and turning it into a custom app (à la Wikimed or PhET) that users can download for free and without any additional data usage beyond that initial download. Easy to install and use, it will help the site reach its intended audience and fulfill its mission of knowledge access.

Project goals[edit]

Note: ZIM - the file format that is used to carry content offline and read by the Kiwix software; Scraper, scraping: the software or act of copying a website, and compressing it into a single ZIM file.

The overarching goal is to launch the Dico des Ados app on Android. This involves:

  • Fixing a number of bugs that currently prevent the mediawiki offliner technology from properly scraping the DdA content;
  • Adding the corresponding recipe to the ZIMfarm servers, ensure monthly updates for the DdA ZIM file and a quality check to ensure that no regression would impact the scraping process;
  • Developing a Playstore presence to host the app;
  • Developing a system for updating the app (engine- and content-wise), preferably in an automated or semi-automated manner.

Project impact[edit]

How will you know if you have met your goals?[edit]

The top goal is to have an application that is up and running on mobile devices. A working Dico ZIM file will also be available for regular download via the Kiwix library, to the benefit of our main francophone deployment partners (Library without Borders, Orange, SAH-OSEO, etc.).

Fixing or improving the performance of the Mediawiki scraper on non-Wikimedia content should impact other contents from the Kiwix catalogue, e.g. Klexikon, Ekopedia, Covid19relay, Cyclowiki, etc. and improve the overall quality of our offering.

Do you have any goals or metrics around participation or content?[edit]

  • Google provides a number of metrics vis-à-vis app usage (acquisition, retention, etc.) which we will readily share, along with the app's rating which we expect to be at least as good as the Kiwix App (currently 4.34).
  • A greater exposure of the Dico should mechanically lead to a greater participation of users, but there currently is no way to directly measure user acquisition. We may run a short survey a year hence with newly registered users to find out their relationship to / knowledge of the app, but this will have to take place outside of the grant's particular timeframe.
  • Considering that the app's audience will not be limited to Switzerland or even France, we will also monitor the share of users from more traditional Kiwix markets (e.g. low-connectivity areas).

Project plan[edit]

Activities[edit]

  • Fix MWoffliner issues related to DdA scraping: 30 person / days
  • Integrate recipe to ZIMfarm, testing: 10 person / days
  • Integrate ZIM file to android app: 15 person / days
  • Develop Playstore presence : 2 person / days
  • Maintenance / updates : 3 person / days

We expect roll out to take place around Q2/2021

Budget[edit]

Type Amount (days) Unit price CHF Cost CHF
Typescript developer (MWoffliner) 30 500.- 15'000.-
C++ Developer (Kiwix-lib, app) 10 600.- 6'000.-
Android developer (app) 10 600.- 6'000.-
Python developer (ZIMfarm) 10 500.- 5'000.-
Project management (8%) 5 600 3'000.-
Total 35'000.-

Community engagement[edit]

Announcements will be made throughout the year on our social media channels (FB, Twitter) as work progresses. Similar annoucements will be made on the local village pumps (Wiktionnaire and Dico);

App publication will be publicized via a Press release;

All code is open-source and Wikimedia CH will be duly accredited as project sponsor on the relevant Github repositories;

The project will be showcased at the various EdTech conferences (on- or offline) we plan to attend this year.

References[edit]