Text & Unicode Tools

When a device has no glyph for a code point it draws a fallback rectangle — the “tofu” box, sometimes with the hex value printed inside it. Older Android builds, smart TVs and set-top browsers are the usual offenders. Windows and iOS cover the mathematical alphabets well; coverage for the more decorative blocks is far thinner, and the newer a character was added to Unicode, the more likely it is to be missing somewhere.

Platforms also filter. Roblox display names reject nearly all non-ASCII input, several games drop the characters silently instead of showing an error, and some sign-up forms validate against a strict allow list. Test in the real destination before you commit to a style, and keep a plain fallback for anywhere you cannot check first.

One emoji on screen is frequently several code points glued together. Skin tone comes from a modifier in the U+1F3FB–U+1F3FF range appended to a base emoji. Multi-person emoji use zero-width joiners (U+200D) between their parts — the family emoji with two adults and two children is seven code points and eleven UTF-16 units. That is why a single backspace sometimes deletes half a person, and why a 280-character limit can be eaten by a dozen emoji.

Older symbols such as ✔ and ☂ have both a text and an emoji presentation. Which one appears depends on a variation selector: U+FE0E forces the monochrome text form, U+FE0F forces the colour emoji. If a symbol you copied shows up black and white when you expected colour, a missing U+FE0F is almost always the reason.

Word counts are tokenisation choices, not facts. Splitting on whitespace treats “state-of-the-art” as one word and “don’t” as one; splitting on non-letters turns them into four and two. Languages written without spaces — Chinese, Japanese, Thai — break the model completely, which is why character count is the honest metric there.

Reading time is arithmetic layered on top: roughly 200–250 words per minute for adult silent reading of general prose, closer to 150 for dense technical material. Keyword density is matched words divided by total words, and anything much above 3% starts reading as spam to humans and crawlers alike. Characters are reported with and without spaces because platforms disagree — an SMS segment is 160 GSM-7 characters, while X weights Latin characters at 1 and CJK at 2.

The symbol browser shows each character’s code point, HTML entity and escape sequence, which matters when the destination is not a plain text box. The entity belongs in markup, the escape belongs in source code, and the raw character belongs in a caption. Copying the wrong one is the most common reason a symbol shows up as literal gibberish in a rendered page.

For interface work, stay with characters that have broad font coverage: arrows at U+2190–U+21FF, box-drawing at U+2500–U+257F, and the classic dingbats around U+2700. Recently added emoji and the ornate decorative blocks are the ones most likely to fail on somebody else’s device.