utf8 or latin1 encoding – German

Question

\usepackage[utf8]{inputenc}

or

\usepackage[latin1]{inputenc}

I write in German if that matters.

Pro and contra?

Answer 1

Short answer: If your editor can use Unicode (and in my opinion any good editor should save files as UTF-8 by default), then use Unicode.

The main advantage of Unicode (and its main encoding UTF-8) is that it makes text files universally readable: Someone writing German umlauts can use the same encoding as someone writing Korean. Prior to that, when obtaining a a text file, there was no way to know which character encoding to apply to it: Latin 1? or Latin 9? or maybe a Russian encoding?

Also for documents that use more than 256 different characters, UTF-8 is very useful. With it you can for example use both "ü" and "ł" in the same text (as I do here).

Both modern (experimental) TeX compilers (LuaTeX and XeTeX) use Unicode/UTF-8 by default as do modern programming languages (i.e. .NET) and operating systems.

Answer 2

If you can: do not use \usepackage[utf8]{inputenc} nor \usepackage[latin1]{inputenc}. Use LuaTeX:

\documentclass{article}
\usepackage{luaotfload}
\usepackage[EU2]{fontenc}
\usepackage{lmodern}
\begin{document}
Das Mädchen ging über die \textbf{Brücke} nach \textit{draußen}.
\end{document}

This will give you access to all modern things (OpenType fonts for example) while keeping most of the backward compatibility.

Wait for TeXlive 2010 (or get the pretest) and you have a decent environment for LuaTeX. Million thanks to the few people who make the LuaLaTeX packages!

If you are able to read german: see the site http://www.luatex.org for more examples (especially on fontspec).

Answer 3

I'm of the idea that, whenever possible, you should stick to utf8 (unicode) as it is finally the solution about character encodings on computer documents. Of course you only have to make sure that your text editor supports and behaves nicely with unicode documents.

In either case, if you have to work with collaborators, you have to make sure that they do also use decent text editors that can handle the appropriate encoding without messing with the special characters. For documents with many collaborators I've learned the hard way that it is better to stick to ascii and use (La)TeX commands to produce special characters.

Answer 4

In short:

• utf8 supports much more characters than latin1.
• utf8 is multilingual and cross-platform spread, latin1 is mainly used on Windows.

Unicode, which is supported by utf8, is international standard and it shall support all languages and shall handle all kinds of writing. utf8 is the widest spread encoding for Unicode and it is backwards-compatible with ASCII (in contrary to other Unicode encodings). In utf8, each character is represented by 1 to 4 bytes. That's why the number of possible symbols is really huge. latin1 uses just one byte, beeing limited to a maximum of 256 characters. I think it uses just 191 of 256 possible.

So, utf8 might be already be favourable if you use just one platform.

utf8 is the standard encoding on Linux and Mac OS X. The number of LaTeX editors supporting utf8 on Windows increased, for instance TeXworks uses it by default.

If there's a possibility that anybody, including you, would open your document on a Mac or on Linux, I recommend to use utf8. This already concerns german umlauts.

Answer 5

One of the advantages of using utf8 for source files is it is friendly towards version control tools. For example, if you have the source in latin-1, then most of latin-1 chars will show up as \267, \320 etc. when diffing.

Answer 6

The background...

Pure ascii is a 7-bit encoding that is only good for english, the german characters is missing.

Latin1 is a 8-bit (1 byte) character encoding that used to be the standard encoding in Germany some 5-10 years ago. And it does support the special german characters. If you have a look at the latin1 page over at wikipedia you quickly realised that it was a little bit messy but it did work.

So to solve this in a nicer way, utf8 was created and since it is using 1-4 bytes to store one character. More or less all special characters from all different countries inclusive the Asian ones can be saved in the same character encoding.

And since all modern OS (Windows, Linux and MacOS) use utf8, you are probably using it right now without even knowing it.

Answer 7

I would also suggest using utf8, but you should definitely check whether all the editors you are going to use (if working with many platforms) actually support it. Although it is 2010, this may still pose a problem. Shame on some developers.

In my personal opinion, using the package ngerman and encoding umlauts via " (i.e. ä = "a) may work better in the long run. It takes a few days to adjust to writing like this, but then your TeX files are ASCII files. With ASCII files, you will never run into any problems.

[I realize that this is not the most modern approach. I myself am quite happy with using utf8 and the venerable vim editor. Your mileage may vary...]