How do I stop Word turning 1/2 into ½?

Press Ctrl+Z right after it happens to undo that one change. To turn it off for good, go to File, Options, Proofing, AutoCorrect Options, and on the AutoFormat As You Type tab clear the box for fractions. It's worth doing for recipes, because Word only converts 1/2, 1/4 and 3/4 and leaves 1/3 as typed, so the list ends up mixing styles.

How do I make a ½-style fraction for 3/16?

Build it from a superscript numerator, the fraction slash ⁄ (U+2044) and subscript digits: ³⁄₁₆. It's plain text, so it copies and pastes anywhere, but fonts don't always match the superscript and subscript sizes. Search won't match it against 3/16 either, which matters for product names people will type into a search box.

Should I write ½, 1/2 or one half in text?

In running prose, spell out a simple fraction: half an hour, one-third of the vote. Use the ½ character with numbers people read at a glance, like 2½ hours in a schedule, a recipe or a table, and fall back to 1/2 only where the text has to stay ASCII. Chicago style also spells out simple fractions in prose.

How do I write ½ in LaTeX?

In text, `\textonehalf` gives the ready-made ½ glyph and works in current LaTeX without extra packages. In math, `\frac{1}{2}` sets a stacked fraction, and `\tfrac{1}{2}` from amsmath keeps it small enough for inline use. For a slanted look on any fraction, the xfrac package's `\sfrac{3}{16}` does it.

Why does parseFloat('9½') return 9?

½ is a single character, and code that expects digits handles it badly. JavaScript's parseFloat('9½') returns 9, dropping the half without an error, and Python's float() rejects it outright. Python's unicodedata.numeric('½') does return 0.5 if you need the value. Normalization has its own trap: NFKC rewrites ½ as 1⁄2 with a fraction slash, so 1½ turns into 11⁄2, which reads like eleven halves.

Why does Alt+189 type ¢ or ╜ instead of ½?

Which one you get depends on your PC's DOS code page, because that's what Alt codes use without a leading zero. A US setup has the box corner ╜ at 189, and a Western European one has ¢ there. Alt+0189 gives ½ on both, and so does Alt+171, where the DOS table keeps its ½.

ASCII 189

One halfBox bottom-right corner, double vertical

ASCII 189 is ½ on Windows and ╜ in old DOS.

½ (One Half) is byte 189 in Windows-1252, Unicode U+00BD. It's the everyday fraction: ½ price, 2½ hours, half shoe sizes. It also turns up in US street addresses. Some houses have fractional numbers like 221½, and USPS addressing standards write them with ASCII digits and a space, 221 1/2. A checkout form that strips non-ASCII characters turns 221½ into 221, a different house, so convert ½ to 1/2 instead of deleting it. Old DOS had ½ too, at 171. In UTF-8 it's C2 BD, which misreads as Â½ on a page decoded as Windows-1252. The HTML entity is ½.

╜ (Double Up, Single Left) is byte 189 in code page 437, the original IBM PC character set, and maps to Unicode U+255C. It's the bottom-right corner of a box with double sides and a single bottom edge: ║ (186) comes down from above and ─ (196) runs in from the left. It completes the set that ╖ (183) starts at the top right, with ╙ (211) at the bottom left. Its entity &boxUl; differs from &boxUL; (╝) only in the case of the last letter, and swapping them draws the all-double corner instead. Code page 850 put the cent sign ¢ on this byte. In UTF-8 it's E2 95 9C.

NumberBox Drawing

Technical Details

Code Example

</>

#include <stdio.h>

int main(void) {
    /* On Windows (code page 1252) byte 189 is ½.
       Modern terminals expect UTF-8, so print the code point, not the byte. */
    printf("\u00BD\n");   /* UTF-8: C2 BD */

    unsigned char b = 189;         /* the raw Windows-1252 byte */
    printf("%d 0x%02X\n", b, b);    /* 189 0xBD */
    return 0;
}

Code Example

</>

#include <stdio.h>

int main(void) {
    /* Byte 189 is ╜ only on a console using code page 437
       (chcp 437 on Windows, or DOSBox). On a UTF-8 terminal the
       lone byte 0xBD is invalid and prints as garbage, often �. */
    putchar(189);

    /* char is signed on x86, so a plain char holding 0xBD is -67.
       Use unsigned char when you compare or index by byte value. */
    char c = (char)189;
    unsigned char u = 189;
    printf("\n%d %d\n", c, u);   /* -67 189 */

    /* Portable: print the Unicode character as UTF-8 instead. */
    printf("\u255C\n");          /* E2 95 9C */
    return 0;
}

How to Type It

Windows: Hold Alt and type 0189 on the numeric keypad, with Num Lock on. The leading zero picks the Windows code page. Alt+171, without the zero, types it too. No numeric keypad? In Word, type 00BD and press Alt+X.
Mac: Press Control+Command+Space to open the Character Viewer, then search for "vulgar fraction one half".
Linux: In GNOME, or any desktop running IBus, press Ctrl+Shift+U, type bd, then Space. With a Compose key set up, press Compose, 1, 2.
Phone: Tap the character at the top of this page to copy it.

How to Type It

Windows: Hold Alt and type 189 on the numeric keypad, with no leading zero. US PCs insert ╜, but PCs set to code page 850, common in Western Europe, insert ¢. No numeric keypad? In Word, type 255C and press Alt+X.
Mac: Press Control+Command+Space to open the Character Viewer, then search for "box drawings up double and left single".
Linux: In GNOME, or any desktop running IBus, press Ctrl+Shift+U, type 255c, then Space.
Phone: Phone keyboards have no key for it. Tap the character at the top of this page to copy it.

Frequently Asked Questions

Press Ctrl+Z right after it happens to undo that one change. To turn it off for good, go to File, Options, Proofing, AutoCorrect Options, and on the AutoFormat As You Type tab clear the box for fractions. It's worth doing for recipes, because Word only converts 1/2, 1/4 and 3/4 and leaves 1/3 as typed, so the list ends up mixing styles.
Build it from a superscript numerator, the fraction slash ⁄ (U+2044) and subscript digits: ³⁄₁₆. It's plain text, so it copies and pastes anywhere, but fonts don't always match the superscript and subscript sizes. Search won't match it against 3/16 either, which matters for product names people will type into a search box.
In running prose, spell out a simple fraction: half an hour, one-third of the vote. Use the ½ character with numbers people read at a glance, like 2½ hours in a schedule, a recipe or a table, and fall back to 1/2 only where the text has to stay ASCII. Chicago style also spells out simple fractions in prose.
In text, `\textonehalf` gives the ready-made ½ glyph and works in current LaTeX without extra packages. In math, `\frac{1}{2}` sets a stacked fraction, and `\tfrac{1}{2}` from amsmath keeps it small enough for inline use. For a slanted look on any fraction, the xfrac package's `\sfrac{3}{16}` does it.
½ is a single character, and code that expects digits handles it badly. JavaScript's parseFloat('9½') returns 9, dropping the half without an error, and Python's float() rejects it outright. Python's unicodedata.numeric('½') does return 0.5 if you need the value. Normalization has its own trap: NFKC rewrites ½ as 1⁄2 with a fraction slash, so 1½ turns into 11⁄2, which reads like eleven halves.
Which one you get depends on your PC's DOS code page, because that's what Alt codes use without a leading zero. A US setup has the box corner ╜ at 189, and a Western European one has ¢ there. Alt+0189 gives ½ on both, and so does Alt+171, where the DOS table keeps its ½.

Frequently Asked Questions

Which one you get depends on your PC's DOS code page, because that's what Alt codes use without a leading zero. A US setup has the box corner ╜ at 189, and a Western European one has ¢ there. Alt+0189 gives ½ on both, and so does Alt+171, where the DOS table keeps its ½.

ASCII 189

One halfBox bottom-right corner, double vertical

ASCII 189 is ½ on Windows and ╜ in old DOS.

Number

Box Drawing

#include <stdio.h>

int main(void) {

/* On Windows (code page 1252) byte 189 is ½.

Modern terminals expect UTF-8, so print the code point, not the byte. */

printf("\u00BD\n"); /* UTF-8: C2 BD */

unsigned char b = 189; /* the raw Windows-1252 byte */

printf("%d 0x%02X\n", b, b); /* 189 0xBD */

return 0;

}

#include <stdio.h>

int main(void) {

/* Byte 189 is ╜ only on a console using code page 437

(chcp 437 on Windows, or DOSBox). On a UTF-8 terminal the

lone byte 0xBD is invalid and prints as garbage, often �. */

putchar(189);

/* char is signed on x86, so a plain char holding 0xBD is -67.

Use unsigned char when you compare or index by byte value. */

char c = (char)189;

unsigned char u = 189;

printf("\n%d %d\n", c, u); /* -67 189 */

/* Portable: print the Unicode character as UTF-8 instead. */

printf("\u255C\n"); /* E2 95 9C */

return 0;

}