Designing a Better Baby Name Search UX and Unicode

A baby-name database initially sounds like a straightforward CRUD problem where you store a name, its meaning, add a few filters, and put a search box on top. That was roughly how Attaullah Siddiqi viewed it when working on Nurturepedia, before observing how people actually search for names in real-world scenarios.

Users do not always type exact spellings. They search by meaning, try different spellings, mix cultural preferences, or look for names from particular origins and religions. Sometimes they know the sound but have no idea how it is spelled, turning a simple database lookup into a fascinating engineering challenge.

Stock photo for illustration only, not from the actual event

The first mistake is treating a baby name as a single string. A useful name record requires significantly more context, separating fields like name and normalizedName. This separation becomes especially useful when datasets contain accents, diacritics, alternate spellings, or characters from different writing systems.

"Search data and presentation data have different jobs."

Attaullah Siddiqi

JavaScript developers eventually encounter the frustrating fact that two strings can look identical while containing different Unicode representations. Characters like é can use a single code point or a base character with a combining accent, which is why String.prototype.normalize() is necessary, though original display values should never be replaced by normalized data.