தமிழ் இணையக் கல்விக்கழகம்Tamil Virtual Academy

The problem this would solve

When the tool is asked for the pure-Tamil equivalent of a borrowed word, it answers from a community-compiled aggregation of purist word lists. That list works. It also has no stated licence, no named editor, and no upstream activity since 2020.

Anyone who inspects our sources will find that, and they would be right to think less of the tool for it. We grade it D and say so in every answer that rests on it, which is honest and is not a solution. A glossary published by the Government of Tamil Nadu is the solution.

What we are asking for

1. A licence statement covering the published TVA archive, so that anyone building on it knows where they stand. Cheap for the institution, and it unblocks more than us.
2. The கலைச்சொல் glossaries in machine-readable form. This is the one that matters. It would let us retire an unlicensed word list and replace it with an authenticated government glossary, with the attribution travelling on every answer.

There is a precedent, and it is theirs. TVA material is already on Tamil Wikisource under CC0, through a collaboration TVA itself documented. Asking an institution to extend terms it has already granted somewhere else is a much smaller request than asking it to invent a policy.

How we treat their material now, unasked

The TVA degree course books are the teaching material this project's grammar tables were checked against. We hold copies. They are deliberately kept out of version control, and only the derived, cited rule tables ship.

The reasoning is the ordinary one: facts about Tamil grammar are not copyrightable and citing a source is scholarship, while redistributing a complete textbook is a different act with a different answer. We did not need permission to do it that way. It seemed like the right way to arrive at a request.

Something we owe them back

While pinning the classical editions we found that the course books quote Nannūl selectively and renumber: their 336, 319 and 136 correspond to 337, 320 and 137 in the Project Madurai edition. Three of our own rule tables carried the course-book numbers and were wrong because of it.

That is not a complaint. It is the kind of thing that is invisible from inside a syllabus and obvious the moment somebody machine-checks every citation, which is a thing this project does by accident of its design. The list is theirs if they want it, along with anything else our citation tests turn up.

thamizh@ief-global.org · we are a nonprofit, the code is Apache-2.0, the service is free, and there is no commercial product built on any of this.