Composite emojis and ZWJ: how do multiple characters become a single shape?

This is why a single figure can be longer for the program than a letter. This has practical significance when copying, deleting, and applying character limits. Let's look at three specific examples of which elements must be kept together.

Let's look at the technologist

The 👩‍💻 sequence consists of the woman, the connector, and the laptop character. The code point of the ZWJ is U+200D; it is not a visible writing mark on its own. The supported sequence can be drawn by the renderer as a single shape. However, inserting a ZWJ between arbitrary characters does not automatically create a new emoji that works everywhere. Source: Unicode Emoji – Emoji ZWJ Sequences

In the third line, the skin tone modifier follows the woman character. It performs a different task than the connector. For example, the 👋🏽 waving hand includes a skin tone, but no ZWJ: choosing a tone alone does not make a sign a ZWJ sequence. The letter pairs of country flags also consist of multiple elements but follow different linking rules.

What does "one character" mean?

When reading, you may perceive as one figure what the program stores as multiple code points. The Unicode text segmentation rules also handle such groups perceived by the user as a single unit. The technical name is grapheme cluster. Source: Unicode Text Segmentation

Therefore, when examining the length limit of an input field, ask three separate questions: how many visible units are there, how many code points do they consist of, and what units does the program count? The label "maximum ten characters" provides little information if the interface counts something different than what the reader expects.

Try it in an empty note: paste the 👩🏽‍💻 figure, then delete it. Does it disappear all at once, or does a person, a laptop, or a tone mark remain? You see the editor's behavior from this. When shortening important text, also check the remaining part.

Why might it fall apart when copying?

Two directions of checking are useful. One is display: does the same complete sequence appear as multiple images in another program? The other is data: did the elements of the sequence survive copying and saving? The two cases may look similar externally, while requiring different fixes.

First, copy the complete symbol into an empty note, then back from there. Then try a simple smiley as well. If only the complex figure changes, specify the concrete sequence and the names of both applications in the bug report. A screenshot shows the appearance, while the pasted text also allows examination of the original characters.

Keep it together when copying and shortening

Do not cut the text in the invisible middle of the figure. When editing, select the entire emoji, and also look at the shortened version. When creating a program, it is worth choosing text processing that handles user-perceived character boundaries, then testing it with specific complex symbols.

The test before publishing should be simple: include a basic smile, a skin-tone hand, and a ZWJ sequence in the test text. Check saving, reloading, and deleting. The guides on emojis appearing as squares and on SMS character limits elaborate on the two related debugging situations.