Strange behaviour in a simple problem



I think the following is safer:

diff(l('p,'c1,'c2,'t1,'t2,'b0,'b1,'b2,'tau,'sigma),'b0);

this way you should have (usually) complete control about what gets 
evaluated and what not. However, there are some functions with special 
evaluation rules, "sum" for instance. I'm not sure if it will always 
produce the desired result...

Martin