Ok here you go. What I did:
- got lingoes compiled dictionaries from http://www.lingoes.net/en/dictionary...&lang=&order=0
- decompiled them with kdictionary-lingoes (lingoes-converter didn't work for me - crashed on records somewhere along the way). Note: the encoding for English is UTF-8 and that for Hebrew UTF-16LE.
- I mass search-replaced = with tab in the resulting txt files
- converted the txt files to StarDict format with pyglossary
This is almost good enough for KOReader. The complaint is that since all tags are removed, formatting is lost, and often space is missing between phonetics, grammatical and definition. I could have used --disable-trim at step 2., but then I would have had to work out all replacements for the tags which are quite nonstandard (like e.g. <H J="rtl">, <Ò>, etc., possibly an artefact of the converter)
For the stock reader, you may ask in the conversion forum for a .dic. I'm curious what comes out with bidi.