User:Fantasticfears/Proposal:TechWorkGroup

From Meta, a Wikimedia project coordination wiki

Many people who still want to contribute to Wikipedia or other Wikimedia projects in China under difficult circumstances. We are glad to see that and appreciate your effort. Though, technology and complexity grows quickly in Wikimedia projects. Hence this document would like to shed light upon future directions for volunteers.

Wikipedia won't exist without its communities. The future contribution of technology needs to be inlined with the community. And the technology will be guard by many different affiliations among Wikimedia communities among the world as well as Chinese community. To be noted, Chinese community consists of people from different regions who might share different political belief. We would always respect their effort as they voluntarily contribute their time and energy with Wikipedia and won't judge their beliefs and worldview.

Technology direction[edit]

We would assume people who want to contribute to Wikipedia with technology want to grow their own skills as well. We would point out possible directions and project descriptions for small group or individuals instead of asking everyone to read obscure and long discussions among different communities.

Questions are appreciated. And this documentation is written in English for better communication with other parties in mind. Of course, do ask me question in Chinese.(可以用中文问问题)

Wikidata[edit]

Kiwix[edit]

  • (moderate) Build a simplified dump for offline content. Language Converter needs to be fully understand before any meaningful work to be done.

Platform evolution[edit]

  • (Difficult) Improve Language Converter system with deep learning models. Language Converter uses a longest matching algorithm to translate Simplified Chinese or Traditional Chinese. Overtime, we have built a large set of dictionaries and it's hard to maintain. With the improvement of DL, it should be possible to improve the current circumstances with less manual auditing and entries maintenance. Due to the scope of such system, further discussions are required and welcome. Funding might be applied to this due to the scope and effort.
    • It's better if a staged rollout.
    • Research on DL model needs to be done before any meaningful work to be started.
    • 100% compatibility on existing policy that we can mixed language variants.
    • Memory consumption should be reasonable in terms of Wikimedia infrastructure.
    • Performance should be reasonable for rendering.
    • Reliable way of auditing and update strategy.

Integration with other projects[edit]

  • (moderate) Write documentation for mw:Multilingual_Templates_and_Modules. This is complicated project and a middle ground because of historical development. But it's promising and seems practical than ever. We would like to invest on it. We need one documentation of on-site review and evaluation of impact.