5

Is there any site, chart, or raw data that shows how much of user have used Unicode (UTF-8) or Western (ISO-8859-1) or Chinese and so.

Or provide rough data of what is used in major number of sites. I want a rough data of major browser setting.

I want to know which character encoding do most user have.

Santosh Kumar
  • 363
  • 3
  • 21
  • 3
    about the "prefer to have" in your question. If by users you mean website users (and not developer users) I confidently think most don't even know what that is, and those who know don't really care as long as it works. On the developer side, UTF-8 is quite standard this days and I can't think of a reason why to use ISO-8859-1 except if you have some old software that requires it. – Osvaldo Mar 18 '12 at 07:18
  • for about normal user, there is a normal setting when they download the browser; what is that? – Santosh Kumar Mar 19 '12 at 03:02
  • My Firefox default is ISO-8859-1 (Latin1). It might be helpful to get a useful answer if you explain what problem you are trying to solve... – Osvaldo Mar 19 '12 at 09:37

3 Answers3

9

Use UTF-8, with proper fonts installed on the client, it universally represents all character sets of all the languages on the planet. Chinese, Cyrillic, Kanji, Arabic, Latin variants, etc. Find the Character Map tool in Windows 7 or its analog in Ubuntu 10+ and have a look at all of them in the various fonts. You can find different fonts are localized to Southeast Asia, Central Europe, India, Western Latin, etc. Modern browsers are supposed to take advantage of this and properly display UTF-8 so you aren't stuck with the Character Set conundrum.

Fiasco Labs
  • 2,472
  • 14
  • 11
  • +1. I'm trying to think of some reason for a modern website not to use UTF-8, and, frankly, I haven't come up with any yet. – Ilmari Karonen Mar 17 '12 at 18:54
  • What is meant by proper font here? – Santosh Kumar May 02 '12 at 05:32
  • You can find different fonts are localized to Southeast Asia, Central Europe, India, Western Latin, etc. So the proper font is the one that has your particular set of squiggles you read daily (Arabic, Chinese, Kanji, Hangul, Hindi, Cyrillic, Western Latin, etc.) While a font may have several different language character sets, I've not come accross one that has 10,000+ characters to fill in all the UTF-8 code points. So the proper one is the one that hits all your needs (the languages you read). – Fiasco Labs May 03 '12 at 02:20
5

What @Fiasco_Labs said as far as which to use. For statistics:

Su'
  • 19,332
  • 3
  • 41
  • 65
2

As Osvaldo comments, it would be helpful to know just why you think you need to know this.

When a user visits a web page, the browser will parse the page using whatever encoding the server tells it to use (via the HTTP Content-Type header or the corresponding HTML <meta> tag).

The only time the browser default encoding matters is when the server doesn't specify the encoding (so that the browser has to guess), or possibly when the server specifies an oddball encoding that the browser doesn't understand (but that's pretty unusual, since browsers tend to know quite a few encodings).

For historical reasons, the default encoding is often ISO-8859-1, since that was the most commonly used encoding back in the very early days of the web, when authors were not so careful about specifying encodings.

Many browsers will also not default to any specific encoding, but will instead try to guess the correct encoding by heuristic analysis if none is explicitly specified. For example, my Firefox default encoding is set simply to "Auto-Detect / Universal"; there are also other auto-detect setting that prefer one encoding (or encodings common in one region / language) over others in ambiguous cases.

But for authors, there's really no reason not to just use UTF-8 these days. Every browser supports it, and, being a Unicode encoding, it contains essentially all characters found in every other encoding.

Ilmari Karonen
  • 9,102
  • 1
  • 24
  • 41
  • 1
    The main reason not to use UTF-8 is storage & compatibility. UTF-8 is a variable-width encoding that uses multi-byte representations for many characters. It's very flexible as it maps to most Unicode characters, but very few applications actually need to support all Unicode code blocks at one time (no font out there has full Unicode support anyway). Even multilingual sites typically only use a single language on each page, so using Latin-1 or equivalent region-specific encodings is more efficient/simpler (many security vulnerabilities are due to mishandling of multibyte charsets). – Lèse majesté Apr 20 '12 at 14:13