Code Points · The lead piece
One code point, and then thirty years of arguing about what to draw
Encoding a character means giving it a number and publishing the table. Agreeing what shape it takes, in every context, in every font, is the part that took decades — and it was settled in fonts, shaping engines and procurement documents rather than in committee.
A chart page shows representative glyphs. The standard’s own note says they are illustrative, not normative.
The Unicode Standard · Devanagari code chartSupporters
Organisations where our research into this history lands.

Section 4 · Procurement
Then somebody had to write it into a purchase order.
Procurement is where a standard becomes a deployment. Kerala and Tamil Nadu specified rendering and language support in documents that are public, and the machines followed.
Three rows of the Devanagari block, sixteen to a row, as the standard prints it
The block →093F is the short i. It is keyed after the consonant it belongs to and drawn to the left of it. A system that draws characters in the order they arrive gets this wrong in every application at once.
0915 is ka. Put it beside 0937 with a virama between them and the two collapse into a conjunct that exists in the font but not in the stored text.
The Unicode Standard · Devanagari block chart
Also worth the time
ShapingThe head-line problem, and the layer built to solve it
The connecting bar above Devanagari means glyphs cannot simply be placed side by side, which is why a shaping engine exists at all.
Read →
StandardsISCII put ten Brahmi-derived scripts on one code table
A single eight-bit table for ten scripts was an elegant answer that Unicode then had to reconcile with everything else.
Read →
ProcurementThe school deployments
Putting one distribution in front of millions of students is a procurement decision with a measurable footprint.
Read →None of this was an argument about desktops. It was an argument about standards, and the standards won.
Four sections, thirteen pieces
Where to go next









