Smart Text Tools

Home › Guides

Why Kruti Dev text turns into gibberish

A legacy font is not an encoding. Here is what is actually stored in those files.

You open an old Hindi document on a new computer and get Hkkjr ,d egku ns'k gSA Nothing is corrupted. The file is doing exactly what it always did; what has gone missing is the font that made it look like Hindi.

What Kruti Dev really is

Kruti Dev is an English font whose glyphs have been redrawn as Devanagari shapes. Press the key for H and the file stores the Roman letter H; the font simply draws भ on the screen. The Hindi exists only in the picture, never in the data.

That worked for twenty years because everyone in a government office had the same font installed. It stops working the moment the text leaves that office: a different computer, a website, a phone, an email, a database search. All of them see Roman letters, because Roman letters are what is there.

What Unicode does differently

Unicode gives every Devanagari character its own number. भ is U+092D everywhere, in every font, on every device. Change the font and the text stays Hindi; it just looks different. You can search it, sort it, read it aloud and send it anywhere.

The two tricky bits in converting

  1. The short i matra. In Kruti Dev it is typed before its consonant because that is where the glyph is drawn. Unicode stores it after, in speech order. Every occurrence has to be moved.
  2. The reph — the little hook for a र that rides on top of the next letter. Kruti Dev puts it after the cluster; Unicode puts it before. It too has to be moved, and past any matras attached to the consonant.

A converter that only swaps characters gets both of these wrong, which is why some online converters produce text that is almost right and reads as nonsense.

What to do with old documents

Typing exams in several states still assess Kruti Dev layouts, which is why the font refuses to die. Working in Unicode and converting at the end is usually less painful than the reverse.

More guides