Talk:Community Resources and Partnerships/India Rapid Project/Creating a modern open source spellchecker for Tamil Language

Initial review questions

Hello Tshrinivasan anf Kaniyam Foundation, thank you for submitting your India Rapid Project application. Your proposal includes a technical component and will be reviewed by the Community Resources & Partnerships and Product & Tech teams at the Wikimedia Foundation. We have a few initial questions to support our review process. Please respond to these questions by 8 July 2025 on this page.

1. Impact on Wikimedia projects and demand from the Wikimedia community:  The India Rapid Project supports projects that aim at improving one or more of Wikimedia projects.

• Could you please provide more details on how the activities described in this proposal, as well as the further development of the spellchecker, will directly impact Wikimedia projects? Please describe both the impact expected directly from the project activities and the potential impact of future activities separately.
• You mentioned creating and editing 1,000 Wikipedia pages, 1,000 Wiktionary pages, and 10,000 Wikisource pages as a result of this project. How do you plan to achieve these outcomes through improvements to the spellchecker? How did you estimate these numbers?
• How have you engaged with the Wikimedia community to assess and confirm the demand for this project? Could you please share any relevant links where you discussed this proposal with the communities?

2. Timeline and maintenance: The India Rapid Project supports distinct projects that ideally should not depend on multiple project requests for ongoing maintenance.

• Could you confirm whether you have already begun work on the spellchecker and whether this proposal aims to improve an existing tool?
• What is your plan for long-term maintenance, and how do you intend to support the project after this project term concludes?
• You mentioned that the spellchecker could be hosted on Wikimedia Cloud and integrated into Tamil Wikimedia projects as a user gadget. Do you plan to implement this integration as part of the current proposal, or is it something you plan to explore later? If you are planning to implement any integration with Wikimedia ecosystem, please describe how.

3. Budget: In your budget request, please specify the number of hours and the hourly rate for each role.

4. Team experience: Could you elaborate on your team by including details about your team’s experience, relevant past projects, and public account handles for developers (e.g., Wikimedia SUL, Gerrit, GitHub, GitLab, or other relevant platforms)?

We look forward to your responses. Please let us know if you have any questions. Best, DSaroyan (WMF) (talk) 10:27, 3 July 2025 (UTC)

Question:

Could you please provide more details on how the activities described in this proposal, as well as the further development of the spellchecker, will directly impact Wikimedia projects? Please describe both the impact expected directly from the project activities and the potential impact of future activities separately.

Answer:

Direct Impact from the Proposed Activities:

The primary goal of this project is to develop a fully functional, open-source Tamil spellchecker, which will have immediate benefits for Tamil-language Wikimedia projects, including Tamil Wikipedia, Wiktionary, and Wikisource. By integrating the spellchecker as a user gadget or tool on Wikimedia platforms (via Wikimedia Cloud), contributors will be able to identify and correct spelling mistakes while editing or creating content. This will lead to:

• Improved content quality: Articles, dictionary entries, and source texts will have fewer spelling errors, making them more reliable and professional.
• Increased editor confidence: Contributors—especially newcomers—will feel more confident writing in Tamil without the fear of making spelling mistakes.
• Inclusive participation: Diaspora users and learners of Tamil who may not have strong spelling proficiency will be more likely to contribute.
• Enabling contributors with assistive tools : New/experienced contributors can focus on creating good content, by reducing the time spent for fixing the spell errors in their and other existing articles. This will help to the goal of “Wikimedia Movement strategy 2030 - Continuous experimentation, technology, and partnerships for content, formats, and devices”

Potential Impact of Future Activities:

Beyond the initial development, future phases of the project can bring deeper and broader benefits:

1. Advanced Features and Integration: With continued development, the tool can evolve to handle more complex grammatical constructions, offer intelligent suggestions, and provide grammar correction. These features can be integrated into the Visual Editor or ContentTranslation tools to support multilingual contributors.
2. Support for Other Wikimedia Projects: The spellchecker can be adapted for other Wikimedia projects like Wikibooks and Wikinews in Tamil, further improving the linguistic quality across the ecosystem.
3. Training and Community Adoption: The project can also serve as a basis for contributor workshops or training programs, educating users on how to write better Tamil and use digital tools for editing Wikipedia.
4. Cross-language Impact: The technical framework and community model can serve as a blueprint for other Indic and low-resource languages, many of which face similar challenges. As an open-source tool, it can inspire or directly support spellchecker development for other Indic languages, especially those that are agglutinative like Kannada or Malayalam.

We previously developed OCR4Wikisource for the Tamil language, which was later extended to support all Indian languages. This significantly improved content quality and productivity across all Indic Wikisource projects. In a similar way, the spellchecker being proposed can also be adapted for other languages—provided there is interest and support from their respective language communities. With a collaborative, open-source approach, this tool has the potential to benefit many under-resourced languages across the Wikimedia movement.

In summary, this project not only addresses a critical gap in Tamil computing but also strengthens the Wikimedia movement’s mission of making knowledge freely accessible and reliable in every language, starting with Tamil.

------------------------------------------

Question:

• You mentioned creating and editing 1,000 Wikipedia pages, 1,000 Wiktionary pages, and 10,000 Wikisource pages as a result of this project. How do you plan to achieve these outcomes through improvements to the spellchecker? How did you estimate these numbers?

Answer:

How the Numbers Were Estimated:

These numbers are conservative projections based on:

• Community activity levels from previous Tamil Wikimedia campaigns.
• The expected increase in contributor confidence and productivity.
• The large backlog of unprocessed or OCRed content on Tamil Wikisource.

We will plan for proofread-a-thon for the wiki articles and pages to achieve these numbers, as part of the implementation of this tool.

------------------------------------------

Question:

How have you engaged with the Wikimedia community to assess and confirm the demand for this project? Could you please share any relevant links where you discussed this proposal with the communities?

Answer:

Through our ongoing discussions and meetings with new contributors, a recurring issue they raised is the lack of a freely available spellchecker for Tamil. Even experienced Wikipedians are aware of this gap. While they often correct spelling errors manually using their own knowledge and experience, this process is time-consuming. As a result, valuable time that could be spent creating new content is instead diverted to fixing existing errors—slowing overall content growth.

We have not yet initiated a formal on-wiki discussion on the Tamil Wikipedia Village Pump, but we are prepared to do so if needed to gather wider community input and endorsement.

------------------------------------------

Question:

Could you confirm whether you have already begun work on the spellchecker and whether this proposal aims to improve an existing tool?

Answer:

Yes, this project is already started. A minimal working version is hosted here - https://iyal.kaniyam.ca/

Code - https://github.com/KaniyamFoundation/iyal-tamil-spellchecker

------------------------------------------

Question:

What is your plan for long-term maintenance, and how do you intend to support the project after this project term concludes?

Answer:

Long-Term Maintenance and Sustainability Plan:

Our vision for this project goes beyond its initial development—we are committed to ensuring its long-term usefulness, adaptability, and community support. The following steps outline our approach to maintaining and growing the project after the grant period concludes:

1. Open-Source Development Model

The spellchecker will be developed under a permissive open-source license, with the code hosted publicly on platform, GitHub. This ensures transparency, encourages collaboration, and allows contributors from both Wikimedia and Tamil open-source communities to submit improvements, fixes, and new features over time.

2. Community Stewardship

We will actively involve the Tamil Wikimedia and developer communities in the project's growth. By building a contributor base during development, we aim to create long-term community ownership. Documentation, contributor guidelines, and training resources will be prepared to support new developers and editors who wish to maintain or extend the tool.

3. Kaniyam Foundation Support

Kaniyam Foundation, with its history of supporting Tamil language technology and Wikimedia projects, will continue to maintain the infrastructure and support ongoing updates. As part of our ongoing training programs, we will encourage students and volunteers to contribute to the tool as part of real-world open-source experience.

4. Wikimedia Integration

By integrating the spellchecker as a gadget or tool in Tamil Wikipedia, Wikisource, and other projects, it becomes a part of the Wikimedia editing environment. This helps ensure continued visibility, usage, and feedback, which can inform iterative improvements.

5. Future Funding and Partnerships

We will explore future funding through Wikimedia Foundation’s Rapid Funds, regional grants, or partnerships with academic institutions and language technology initiatives. The project’s modular design will allow for easy scaling and extension based on available resources.

In short, this project is designed not as a one-time effort, but as a sustainable digital asset for the Tamil language and Wikimedia ecosystem—driven by community, supported by institutions, and maintained through open collaboration.

------------------------------------------

Question:

You mentioned that the spellchecker could be hosted on Wikimedia Cloud and integrated into Tamil Wikimedia projects as a user gadget. Do you plan to implement this integration as part of the current proposal, or is it something you plan to explore later? If you are planning to implement any integration with Wikimedia ecosystem, please describe how.

Answer:

Planned Integration During This Project:

1. Wikimedia Cloud Hosting: We intend to host the spellchecker service on Wikimedia Cloud Services (Toolforge or Cloud VPS), making it accessible to Wikimedia contributors without requiring local installation. This will ensure high availability and seamless access for Tamil Wikimedia projects.
2. User Gadget for Tamil Wikipedia and Wikisource: We will create a simple JavaScript-based user gadget that allows editors to check and highlight spelling errors while editing content on Tamil Wikipedia and Wikisource. This gadget will interact with the hosted spellchecking API and display real-time suggestions.
3. Testing and Feedback Loop: As part of the integration, we will test the gadget with active contributors, gather usability feedback, and improve its functionality to better fit the editing workflows.

------------------------------------------

Question:

3. Budget: In your budget request, please specify the number of hours and the hourly rate for each role.

Answer:

Added the details in the sheet below

https://docs.google.com/spreadsheets/d/1jRxd3PKv-q-gOc49PKKGr_bUImI-6I1W/edit?gid=1454251968#gid=1454251968

------------------------------------------

Question:

4. Team experience: Could you elaborate on your team by including details about your team’s experience, relevant past projects, and public account handles for developers (e.g., Wikimedia SUL, Gerrit, GitHub, GitLab, or other relevant platforms)?

1.

Backend Developer - Shrinivasan

Developer/Devops engineer

Blog - https://goinggnu.wordpress.com

Github - https://github.com/tshrinivasan/

Wiki Username - Tshrinivasan

Shrinivasan is one of the co-founders of the Kaniyam Foundation, an organization dedicated to promoting Free and Open Source Software (FOSS) and open knowledge for over 14 years. Through our online magazine Kaniyam.com, we regularly publish technical articles related to FOSS. Additionally, our ebook platform FreeTamilEbooks.com hosts over 900+ Creative Commons-licensed Tamil ebooks spanning a wide range of genres.

Shrinivasan is an active contributor to Tamil Wikipedia and the broader Wikimedia ecosystem through programming, community training, and organizing hackathons and meetups. One of his significant contributions to Indic Wikisource projects is OCR4WikiSource, a tool developed between 2015 and 2019. This tool has facilitated the OCR (Optical Character Recognition) of over one million pages across various Indian-language Wikisource projects. It automates the OCR process for PDFs on Wikisource, enabling contributors to efficiently extract and proofread texts.

2.

UI Developer - Syed Jafer

Full Stack Developer

Website - https://parottasalna.com/

Linkedin - https://www.linkedin.com/in/syedjaferk/

Github - https://github.com/syedjaferk/

Wiki Username - Iamsyedjaferk

3.

Tester1 - Hariharan

Web developer / Tester -

Linkedin - https://www.linkedin.com/in/hariharanu/

Git - https://github.com/HariharanUmapathi

Site - https://hariharan-u.web.app

Wiki Username - Hariharanumapathi

Hariharan is the organizer of Kanchi Linux Users Group. He conducts weekly meetups on Free/Open Source software in Tamil language.

4.

Tester2 - Lenin Gurusamy

Ebook Publisher at https://FreeTamilEbooks.com

Audio book publisher at Olipedia - https://www.youtube.com/@olipedia/videos

Git - https://github.com/gurulenin

Wiki Username - Guruleninn

5.

Data Curator - Boopalan

Developer

Git - https://gitlab.com/boopalan-dev/

Lead developer of Tolkapy - A Tamil language grammar rule engine in Python.

Wiki Username - Boopalan28012003

6.

Data Validator - Kalaiarasan

Developer

Blog - https://kalaiarasan3.wordpress.com

Wiki Username - Kalaiarasanpandi

7.

Tamil Grammar / Linguist Expert - SME - Sathyaraj

DR.T.SATHIYARAJ M.A. (Tamil), M.A. (Linguistics), MPhil, Ph.D., NET., TPT., CCMT., DIPLOMA IN SANSKRIT, DOM, is an Assistant Professor, Department of Tamil in SriKrishna Adithya Arts and Science College. He has presented 86 papers in various Conferences, International Seminars, National Seminars, Journals and chapters.

ResearchGate: https://www.researchgate.net/profile/Sathiyaraj_Thangasamy

Googlescholar: https://scholar.google.co.in/citations?user=TR8U4lUAAAAJ&hl=en

ORCID : https://orcid.org/0000-0002-2272-3117

Wiki Username - Neyakkoo

8.

Machine Learning expert - Nithya Duraisamy

Developer - Author of 14 free ebooks including Machine Learning and Deep Learning in Tamil language.

Ebooks - https://freetamilebooks.com/authors/nithyaduraisamy/

Git -

https://github.com/nithyadurai87/

Wiki Username - Nithyaduraisamy87

Wikisource Bot -

Nithyaduraisamy87BOT

Tshrinivasan (talk) 23:05, 8 July 2025 (UTC)