Do not translate the actual markers. HTML consists of tags which indicate the start and end of a section of text. This text could be a heading, a paragraph, a hyperlink or just a piece of text to display in bold:

In some cases, it might be necessary to reorder the tags. If tags surround some text, it means that the translation representing the text inside the tags should probably also be surrounded with the same tags.

Some inline tags can represent something that should be positioned in the translation, such as an image or a footnote. Take care where you put this in the translation to have the same effect as the original text.

Whenever you work with tags, always ensure that opening and closing tags go together in pairs, and that the closing tags always follow the opening tags.

Although as a translator you should only change the text of the program, you do have complete control over the HTML. This can be used to your advantage if needed. If for example you were translating a manual that referred to an image of the application then you would most likely want to have an image in the users language, rather then explaining to them in their language using an English picture.

Because of the work involved, the translation teams will often not change images until much later in their efforts.

To change the image first you will need to create an image. You can use ksnapshot or gnome-screenshot or Gimp to create screenshots. Remember to keep the look consistent, i.e. try to use the same theme across all images. Ideally your picture should be the same size as the one it is replacing.

Lastly depending on the application you will need to place the file in the correct place so that it will be shown with your translations. What usually happens is that images are first sought in the language specific directory and if they are not found then they use the English version.