Maybe a little off-the-mark, but from the rec(?:ei)|(?:ie)ving end, i think i benefitted most not from 'case study' type advice ( like "it would go faster if..." or "learn your own" ), but from advice that gave me the independence to trouble-shoot and experiment myself, at which point i could then return with questions like "how can i make this faster?" or "How would you do this differently?".

By this is mean being shown how to properly read error messages and warnings, using perldoc and cpan effectively, and the debugger. As well as other community resources like perlmongers ( and the Monastery of course! ), which give you the ability to tap into and see what other people are doing and how they do it, but i guess that comes back to the "learn your own" part, where you start to read|analyse|get inspiration from other peoples code...

I guess it is a very personal thing and everyone goes about the learning process differently, which is what makes mentoring so difficult, but at least for me, being given that independence to experiment ( and the tools to test those experiments ( can you tell i am a scientist...) ) was the biggest catalyst.

Though now i think about it, I think i also value having places i can go for help when my experimenting doesn't pay off, or when i just want to run an idea past people, so i guess it is a balance between giving your tutee freedom, whilst remaining a ( trusted ) presence.