This was the title of a talk that I gave at the " Future Trends in Hypercomputation " Workshop back in 2006. A revised version of the original paper has been posted to arXiv.org: http://arxiv.org/abs/0910.2859
Bruno Scarpellini was kind enough to point out to me a minor misprint that he discovered. In particular, on page 178 of line 11 froom above, that is, in clause (i) of definition 9.3.4, one has to replace H(z(\lambda)) dy/d\lambda with H(z(\lambda)) dz/d\lambda, that is, one has to replace dy with dz.
Last year I wrote a "paper" on hypercomputation, but since it is not quite ready, I have posted it to the arXives. I believe the "paper" contains some interesting ideas. Anybody willing to have a look at it, can fetch an electronic copy from the following URL: http://arxiv.org/abs/0910.1494