12 August 2026
Four catalogues, not one
The cheap way to build a quotes app in several languages is to hold one list and translate it. One row, four strings, one set of tags, one attribution to check. Everything downstream gets simpler: the search index, the widget, the share card, the lot.
We did not do that, and the reason is that the cheap version tells a lie about the material.
A Tamil couplet from the Thirukkural is not an English aphorism that happens to be in Tamil. It is two lines of seven feet, written to be memorised, sitting seventh in a chapter of exactly ten, in a book of 133 such chapters, with the chapters on either side of it doing half the work of its meaning. Kural 481 says that a crow beats an owl in daylight. Rendered as a standalone English sentence it reads like a nature fact. Read where it sits, as the first couplet of the chapter on knowing the right moment, immediately after the chapter on knowing your own strength, it is a filed observation about timing beating force, and the bird that loses is the stronger one.
None of that survives translation into a list.
What four catalogues actually costs
Today the shelves stand at 56,391 English, 3,149 Hindi, 2,169 Tamil and 940 Sanskrit. Each one was built from its own sources: the Kural and the Eighteen Lesser Texts for the Tamil, the Kabir Granthavali and Rahim's dohe for the Hindi, the Gita and the subhashita collections for the Sanskrit. There is no row anywhere that is a translation of another row.
The costs are real. Four catalogues means four sets of sources to establish rights on, four transliteration schemes, four typefaces drawn for four scripts, and four times the reading. It means the Sanskrit shelf is 940 rows against English's 56,391, and that gap sits on our own marketing page where anybody can see it.
It also means the app cannot cheat. When you switch language, you do not get the same thought in another script. You get a different verse, by a different writer, from a different century, and the English underneath it is there to help you read it rather than the other way round.
The number that looks worst
940 is the number people ask about. Sanskrit is the slowest shelf to grow because it is the one where a wrong reading is easiest to produce and hardest to notice. Every line on it was read by a person before it went in, and a good few were thrown out at that stage for being fragments of a chapter heading rather than a verse.
We would rather have the small honest shelf than a large one nobody has read.