1: \begin{abstract}

2: We make the case that the so called ``global decision problems''

3: should not be investigated solely for groups described by finite presentations.

4: We propose to use descriptions that be algorithms that perform some

5: given tasks, and that encode the considered groups. We motivate this

6: by establishing undecidability results for groups described by recursive

7: presentations, strong enough to prevent an interesting theory of decision

8: problems based on generic recursive presentations to be developed.

9: More importantly, we give an algorithmic characterization of finitely

10: presented groups, in terms of existence of a ``marked quotient algorithm''

11: which recognizes the quotients of the considered group. This new point

12: of view leads us to proposing several open questions and directions

13: of research, and much of this paper consists in exposing problems

14: that arise from our first results. Finally, note that we set our study

15: in the category of marked groups, we explain why this is beneficial,

16: and give open questions that arise from the study of decision problems

17: for marked groups. 

18: \end{abstract}

19: