There seems to be a rather interesting development in this issue during this past week. But first, a little bit of a very, very short history on this.
The whole "holy grail" of superconductivity (and condensed matter physics, in general) is finding the "glue" that causes pairing in the high-Tc superconductor family. This glue will be the direct mechanism that causes this phenomenon in these material, very much like phonon being the pairing glue of conventional superconductors.
Currently, there are two competing scenarios for the possible candidates for this glue: phonons and magnetic interactions. While these two scenarios have been floating around for a while as the possible mechanism for high-Tc superconductors, they both came to a head-on clash with the publications of 3 papers in the same year. The papers by Kaminsky et al. and Johnson et al. support the idea that the magnetic channel is responsible for the electronic properties in these superconductors and thus, is the responsible mechanism for superconductivity. On the other hand, the paper by Lanzara et al. argued for the phonon mechanism in the same material. What is interesting here is that all three papers essentially are looking at the same type of experimental results! All of them are studies using angle-resolved photoemission spectroscopy (ARPES), and all of them are looking in particular at the "kink" feature along the nodal direction of the reciprocal space of the material, with varying doping! So what you have here is the same type of experimental results, with 2 different interpretations.
While there have been many papers arguing for both phonons and magnetic interactions since then, two recent papers have appeared in Nature that made definite claims for the phonons being the responsible glue. The first one, published in 2004, was a new ARPES study of the isotope effect in the bismuth cuprate superconductor. Now, while the isotope effect has been found to influence the value of the critical temperature Tc in conventional superconductors, such an observation isn't found in high-Tc superconductors, which was the initial impetus for many to drop the phonon mechanism in these material. However, this paper showed that there is a large effect on the electronic structure seen in the ARPES result when an isotope substitution was made (i.e. changing O16 to O18). This was seen in the high energy broad "hump" in the spectrum.
A new rebuttal of that paper was just published this week that repeated the same measurement (and more), and found no such effect. While they do not claim to dispute the phonon picture, they certainly threw doubt into the experimental results of the earlier paper.
The second paper that also originally supported the phonon mechanism was based on results from scanning tunneling microscopy (STM). Here, again, the energy scale of the "dip-hump" feature that is common in tunneling results was examined as a function of substitution between O16 and O18 isotopes. The authors argue that the "mode frequency" changes dramatically with such substitution, and such dependence argues for the phonons as the source that this mode.
This interpretation too is disputed. A new paper has argued that the tunneling results that was seen was primarily due to the inelastic tunneling effect. It is argued that the paper was probing the excitation of the apical oxygen that resides in the insulating later, and not the conducting Cu-O plane where superconductivity is believed to occur.
Moral of the story: the phonon picture is far from having any convincing results that would make this as the pairing glue.
 Kaminsky et al., PRL 86, 1070 (2001).
 Johnson et al., PRL 87, 177007 (2001).
 Lanzara et al., Nature 412, 510 (2001).
 Gweon et al., Nature 430, 187 (2004).
 Douglas et al., Nature 446, E5 (2007).
 Lee et al., Nautre 442, 546 (2006).
 Hwang et al., Nature 446, E3 (2007).