Grants talk:Project/MSIG/Wiki Term Base

Feedback for improvement

Dear @ عباد ديرانية Kudos to the project team on designing a clear and detailed experimental project. The project targets a significant issue on Arabic Wikipedia - standardizing vocabulary usage - which is important for consistency, and has been a cause for numerous challenges the community has had to grapple with over time.  It is also good to see that the team has developed a good subject matter expertise partnership with the WMF Language Team, Columbia University’s Natural Language Processing group, and the Arabic Language Academy. This shows a commitment to leveraging diverse expertise and resources.

Feedback for improvement:

• Enhance community engagement strategies: While the project mentions engaging the community at various stages, consider implementing structured mechanisms such as regular updates through dedicated channels (e.g., newsletters, forums) and hosting interactive workshops or webinars to solicit ongoing feedback and foster a sense of ownership among community members. Doing this will ensure that there are diverse perspectives and voices, especially from underrepresented communities within the Arabic Wikipedia ecosystem. This can enhance the relevance and effectiveness of the tool generally.
• It is important for the team to clarify sustainability plans following this experimental stage. While the project timeline is well-defined, it would be beneficial to elaborate on long-term sustainability strategies beyond the grant period. Consider plans for maintenance, updates, and community-driven contributions to ensure the tool remains relevant and functional over time.
• It is strongly recommended that the team prioritizes usability and accessibility. As the tool aims to streamline vocabulary standardization and translation processes, prioritize user experience design and ensure accessibility for users with diverse backgrounds and technical proficiencies. Conduct usability testing and iterate based on feedback.

YPam (WMF) (talk) 19:11, 12 June 2024 (UTC)

Thank you so much Yop for the great feedback from you and the team!

• Considering that we chose an agile product development methodology for this project, we do aim to involve users at every step along the way. We will do our best to complement this with a community engagement strategy, while also taking in mind our projected timeline for the tool development.
• Sustainability is a great point. I just edited the grant draft to include a key feature: figuring out community contributions through the tool. This is also closely tied to Nada's comments below - we do hope to integrate the tool into Wikidata and/or Wiktionary, either of which could be pathways for user to contribute directly to the data underlying the tool on the longterm. We still need to figure out the technicalities of how this may work in practice, however.
• Also part of our agile development methodology, and usability testing is very core to our plans!

Abbad (talk) 16:56, 16 June 2024 (UTC)

Questions

Hello,

Thank you @عباد ديرانية for putting the effort on this project.

I have a few questions:

If I understand correctly, you are creating an online dictionary. How will it differ from the many existing online dictionaries? Beyond the wiki integration process, will there be additional benefits for Arabic contributors?

How will you ensure that copyright issues are not violated when using existing dictionaries?

Have you thought about integrating Arabic Wikitionary in the tool?Nada.FA (talk) 07:38, 14 June 2024 (UTC)

Thank you for your questions, Nada! Great to hear from you.

• Partly correct! Our main deliverable for the current stage is similar to a dictionary, but it's in fact a dictionary database. We aim to collect data from hundreds of dictionaries in one place, allowing Arabic Wikipedia users to immediately access a summary of the lexical data from across all of them, while they are editing Wikipedia pages. For example: when searching the word 'Polymerization', you can see a ranking of the corresponding Arabic terms by their frequency in dictionaries, e.g. term number one came up in 70 dictionaries, the next came up in 40, etc.
• We will not make the entire lexical entries and text accessible to the user. Instead, we are using natural language tools to process the text and aggregate dictionaries using the same terms. We will then cite these dictinoaries through the tool, similarly to how Wikipedia articles would cite them without quoting the exact text.
• We are thinking about both Wiktionary and Wikidata integration, as we do hope that the project will produce an open database that Arabic Wikipedia users can update. It would be ideal if this can be stored on Wikidata, where the list of dictionaries we aim to use is already stored. Still figuring out the exact technical details, though!

Abbad (talk) 16:47, 16 June 2024 (UTC)