always be very careful and make sure you will be accredited for the work you have done - even if you are just doing what you are told.

my supervisor published work done by me in collaboration with someone elses, she got her name on it and I got a thankyou at the end!

End result - not a listed publication for me!

Lesson make sure any work you hand over for publication includes you!!

cont....
Start by focusing every experiment on where it would fit into a paper - it stops you doing random experiments that mean nothing.
If you want to go a completely different direction, make sure there is sound reasoning behind it.
Remember if you go way off topic, your supervisor may not have the knowledge to help you later when you need it- but if you stick too close to current publications, chances are someone is already doing it! - take a look at the numerous threads from people who were out published!
Its a game, but pick the right direction early on and you'll make you life a lot easier
Good Luck

I had the same problem, didn't think it was going anywhere imaginative.
No.1 - its your project - you decide - supervisors provide help and support, but after year one it should be up to you.
The best thing I always advise is time spent planning well now will save you hours/ weeks/ months later
Start by looking at the current publications you are going along with then think - where could I publish in all of this - try incorporate what you have already done- this could take you a month but its well sorth it!

most scientists will talk - alot! about what they do. Try take key points from what they say and relate them work you do, and form questions that are a point for discussion.
Mentioning funding wishes is fine, just make sure they are realistic! it might be best to ask 'I was interested in getting this funding to help, I was thinking about going about it this way. What do you think?' a straight talking supervisor will be encouraged you are thinking and realistic.
A question in interview I once heard was:
'its friday afternoon, you have a paper due for submission on monday morning - but all your friends are going to the pub- what would you do?'
everyone answers 'I'd work till all the way through and submit in time(probably die of a heart attack 3 seconds after submitting)'
The guy who gets the job is the one who says
'I'd go to the pub have a couple- get the stresses out and wake up on saturday refreshed and get it done'
Supervisors usually prefer honesty and realistic goals.

As frustrating as it is, this is the way it works. Everyone is racing to publish their data first. It happened to me and it happened to my supervisor and it can be really frustrating.
The good news is that it doesn't stop you writing your thesis, what I would probalby do is to include their reference in your discussion as backing up the data that you have found- despite everything it is always nice to know you were right!
As for publications the option is to try find SOMETHING you have done that they didn't and use that as your angle to publish, so you show the same data but you might have looked at a very slightly different aspect of the work - I would examine their paper very closely and look for imperfections you can exploit.
If all that fails, it really isn't the end of the world - write your thesis and find a job with someone who publishes loads and learn from the experience.
Main thing is to get the thesis done