A code point like 邉 (U+9089) has 32 Ideographic Variation Sequences registered in the Unicode IVD — the most of any character. But registration only says “this combination is meaningful”; whether you actually see a different glyph depends entirely on the font.

Open one of the most comprehensive Japanese fonts on most operating systems (e.g. Yu Mincho, MS Mincho) and you will find that most of the 32 variants render identically to the default. The font simply doesn't carry distinct outlines for every registered IVS.

The IVD is split into collections, each maintained by a different organisation with a different scope. A font typically targets one collection — not the whole IVD.

A single character can be registered in multiple collections under different VS numbers. For 邉:

A font that knows only about Adobe-Japan1 will leave the Moji_Joho range mapped to the default glyph. A font that knows only about Moji_Joho will leave the Adobe-Japan1 range alone. To render every variant correctly you need a font — or a combination of fonts — that covers every collection the character appears in.

Even within a single collection, some IVS entries are deliberately registered as identical to the default glyph. The font marks them in its cmap format-14 table with None as the glyph reference, meaning “use the base character's default glyph here.”

This isn't a font bug or omission — it's the IVD's way of giving a stable identifier to a variant whose visual shape matches the default. For 邉:

The Unicode Viewer detail panel marks these three states with three different badge colours: blue for distinct IVS glyphs, cyan for “= default” entries, and grey for entries the font doesn't register at all.

To show every registered variant honestly, this site bundles a single Web font assembled from three upstream sources, each filling a role the others can't:

Why three? Because each is the only practical option for a particular gap:

• Jigmo is the only freely-licensed font with a full IVD 2025-07-14 implementation. Using it as the single source for IVS means base + variant glyphs share one type designer's style, so the side-by-side comparison in the detail panel highlights structural differences instead of typographic drift.
• Source Han Serif ships a Standardized Variation Sequence (SVS) table that Jigmo lacks — without it, the centred punctuation forms (、︁ 。︁ used in vertical Japanese writing) would have no glyph at all.
• IPAmj明朝 covers a handful of Moji_Joho-specific SVS entries that Source Han Serif doesn't register.

Where Source Han Serif provides the variant glyph for an SVS, the merge also overrides the base cmap entry to point at Source Han Serif's default glyph, so the bare character and its variant render in the same typeface. Without that step, comparing 。 with 。︁ would mix two designers' hands and obscure the structural difference.

If you build a tool, document, or website that relies on IVS for legal records, kanji name handling, or historical text, font choice is not cosmetic — it determines whether the variants are visible at all.

• Verify which IVD collections your target font covers. Adobe's site, the IPA mojikiban downloads, and Jigmo's README all publish their coverage.
• If a variant looks identical to the default, check whether the font actually has a distinct glyph or whether the IVD entry is a “= default” registration. The two situations need different fixes.
• For web delivery, consider chunked WOFF2 with unicode-range so the user only downloads what they actually render. This site uses 121 chunks of 256 code points each, ~14 MB total but typically <1 MB per page view.