[QUOTE=ipse;2823354I can't figure out how to get the dictionary work....

[/QUOTE]
If there is digital text, you don't need tesseract. If the text is scanned only, as in images of text, you will need Tesseract traineddata-files in the
koreader/data/tesseract/.
Sometimes people start trying out with a non-verified Stardict dictionary. When you install Stardict for windows, it comes with an editor which can verify your dictionary.
Sometimes people forget to decompress the dict.dz-file to an dict-file.
Sometimes people forget to place all three files, dict-, ifo- and idx-files in the koreader/data/dict directory.
Anyway,
here are some working dictionaries.