I personally find it rather impossible to work an “hard” projects - which are often the subject of delayed start - for only 30 minutes. Even the prospect of not having at least three hours for such a start of a new big job stops me from working on it. Once I am involved with a subject I do not find it hard to stick with it for hours on end.

So here is my strategy:
Split the todo-list into “small and useful” (like: fiddle with website, do bookkeeping, …) and “large and hairy” (like: write a specification, write a large program, …). Of course each of these lists also needs to be sorted according to priorities (deadlines etc.)
Now start with the item that you least like and that is most pressing, given the time available for it today. Thus, if you only have 30 minutes left, don’t start on any big, hairy item. If you do have ample time, however, start with the one you like least.
When you get tired of it, do a small useful one and return to the current big one. When you have done all the small useful ones, switch back and forth between two big ones. Promise yourself the pleasure of working on a nice big one (develop a new program with new tools) as a reward for having finished the necessary and pressing ones (write the controversial strategy paper for a big project).

I find that I can get involved even with stupid administrative jobs once I have started them if the reward is available. The only real difficult show-stopper is difficult relationships with people for who you write or develop. I have not really found a solution for that problem. I try to avoid working on such projects for such customers.