Why Gujarati is not a translation of Sitegride

There is a familiar shape to business software built in India: it is designed in English, sold in English, and then, once there is demand, a language file is added. The buttons change. The headings change. And then you open a real screen and find the date format is still American, half the error messages are untouched, and the one place where a decision actually gets made is in a second language.

This matters more here than in most categories, because of who is holding the phone.

Who is actually entering the data

The owner is often comfortable in English. The person recording eighty percent of the entries is not. A site manager marking hazri, a supervisor receiving a lorry, a store keeper checking a delivery challan: these are the people the record depends on, and they are quick, precise and confident in Gujarati.

Give them a screen that is Gujarati at the top and English at the point of decision, and you have not built a bilingual product. You have built an English product with a Gujarati wrapper, and it will be used slowly, then reluctantly, then not at all.

What "equal" actually requires

Treating two languages as equals costs more than translating, and it is a different set of decisions:

• Both languages get a real typeface for the same role, not one good font and one fallback. Latin and Gujarati are set in faces chosen to sit together at the same weight.
• Layout is designed for the longer of the two strings. Gujarati often runs longer, and a button that fits in English and clips in Gujarati is an English-first product.
• Numbers stay in the Indian system (lakh and crore, and the 12,34,567 grouping) in both languages, because that is how the business thinks regardless of script.
• Errors, empty states and confirmations are translated with the same care as headings. These are the moments where a user is already unsure.
• The language switch is available everywhere and remembered, because a household or an office is rarely all one thing.

The vocabulary problem

The hardest part is not grammar; it is deciding which word a site actually uses. Formal Gujarati has a translation for "attendance", but the site says hazri. It has one for "contractor", but the site distinguishes mistry from the sub-contractor who bills you. Choosing the formal word makes the software correct and unusable.

So the vocabulary comes from sites, not from a dictionary, and where a term is genuinely English on site, it stays English in the Gujarati interface, transliterated the way people say it. A product that insists on purity in a place where the trade does not is just another form of not listening.

Software people cannot read is not simpler. It is just slower, and then it is abandoned.

Why this pays for itself

Every argument for bilingual-by-default eventually comes back to one thing: the record is only as good as the entry, and the entry is made by whoever is standing at the gate. Anything that makes them slower makes the owner’s numbers older.

The language of the interface is not a courtesy. It is the difference between a system that is filled in and one that is filled in on Saturday, from memory, by somebody who was not there.