[ftarlao]

Yes it works, it finds out the 'shape' of the surrounding text and it writes a regex using lookarounds operators which does the work right. So, yes, it builds an extractor which is able to extract only "simon" and "mark"