Unicode variation selectors (2007): essential tool or overreaching fix?

The debate, in plain sight

One of Unicode's most divisive features is the variation selector. Some call them "pseudo-coding"; others want to throw a variation selector at every new encoding problem. The practical middle ground? They are indispensable for controlling contextual glyph forms in complex scripts like Mongolian and Phags-pa, but they are a blunt — and often inappropriate — instrument for marking stylistic or epigraphic variants, and they should never become private glyph IDs. Scholars who insist that encoded text must be an exact facsimile of an inscription raise a heartfelt, real concern. But encoded text and photographic facsimiles serve different purposes — both are needed. Which do you want your plain-text corpus to be: a manuscript photo, or a searchable encoding?

How they work (the short version)

Variation Selectors are discrete Unicode characters used to form variation sequences: a single base character followed by one selector. There are 256 generic selectors (FE00–FE0F for VS1–VS16 and E0100–E01EF for VS17–VS256) and three Mongolian Free Variation Selectors (180B–180D, FVS1–FVS3). Important technical constraints apply: the base must not be decomposable or a combining character, otherwise normalization can change what the selector attaches to — a gotcha that bit the early mathematical variation sequences.

Standards matter

Crucially, variation selectors are not a free-for-all. Only variation sequences that Unicode itself defines — standardized variants — are meant to be recognized by conformant implementations. You can locally pair A + VS16 to mean whatever you like, but you shouldn't expect toolchains or fonts to honor it. It has been reported that Microsoft Vista supported some variation sequences not defined by Unicode, an eyebrow-raising exception to the rule that illustrates how messy real-world support can get.

Looking ahead

With more historic scripts being encoded, calls to use variation selectors to capture myriad stylistic letterforms are growing louder. The author of the original analysis predicts a proliferation of such sequences. But is that the right direction? Higher-level markup or font-level distinctions often solve the epigraphic problems more cleanly than bloating the character repertoire. The debate is part technical, part cultural — and it will shape how the digital humanities remember the past.

Sources: babelstone.co.uk, Lobsters