Conquering Character Encoding Chaos With GNU Recode

December 4, 2008
By

Carla Schroder

In the beginning were C and C++, and hosts of other computer programming languages. These are all based on ASCII (American Standard Code for Information Interchange), which as the name implies is based on the English alphabet. Which wouldn't be an issue except there are lot of other humans in the world, and they don't use the English alphabet.

So along came Unicode to the rescue. Unicode provides a framework for all alphabets of the world to be represented on computers. UTF-8 is the most popular Unicode implementation because it preserves backwards compatibility with ASCII. Which is all fun to know, but what good is it when you're looking at piles of computer files that need to converted from ISO-8859-1 (Latin-1, Western European) into whatever encoding you prefer? Naturally, there are a number of utilities just for this task.

GNU Recode supports over 150 character sets, and converts just about anything to anything. For example, there are still users of legacy Linux systems that still run ISO-8859-1. Recode will convert these to nice modern UTF-8, like this:

$ recode UTF-8 recode-test.txt

Check out the GNU Recode Manual for instructions.

That's fast and easy enough, but there's one more job- converting the filename. The convmv command is just the tool for this job. This example converts all the ISO-8859-1 filenames in the files/ directory to UTF-8:

Email Article Print Article Comment on this article Share Articles Digg del.icio.us Newsvine Facebook Google LinkedIn MySpace Reddit Slashdot StumbleUpon Technorati Twitter Windows Live YahooBuzz FriendFeed

$ convmv -f iso-8859-1 -t utf8 --notest files/ convmv run without the --notest option does a dry-run without changing anything, which is probably a wise thing to do first.

Maybe you have a file that you don't know what the encoding is. Upload your file to this online tool and it will tell you. You can even do file conversions here.

Resources

The subject of character encoding is huge and bewildering, especially for us dinosaurs from the typewriter era. By golly, when you hit a typewriter key it came out the same way every single time. Wikipedia has a number of excellent introductory articles:

Unicode
UTF-8
ISO/IEC 8859

Linux Clone of Microsoft's Silverlight in Beta

Red Hat Spacewalk Expands Linux Management

Linux Moves From Grazing to Gorging at the Unix Buffet

Red Hat Fedora Claims It's the Leader in Linux

Reducing IT Costs with Linux

Fedora 10: the GNU/Linux Desktop Steps Forward

Red Hat Fedora Claims its the Top Dog in Linux

This article was first published on LinuxPlanet.com.

Tags: Linux, programming, IT, Linux, IEC