The dodo is, as we know, a little bit extinct. And while there have been some sightings in the 1990s on Mauritius of a strange bird, the island is significantly small enough that these can likely be discounted. So if the bird were to be revived, it'd have to be done via genetic engineering.
The chance of this is compounded by a few problems. The first is that we don't actually have much dodo "material" (for want of a better word) around; in fact, there's just a single head and foot in Oxford, and a single skeleton in Mauritius. The second is that it's somewhat a nih-nor situation anyway: the dodo DNA that we do have is decomposed to such an extent that creating a living dodo from it is nigh on impossible. As the Sunday Times science correspondant put it:
Once scientists have worked out the key genes that made the dodo unique, they could then create genetically engineered DNA to put into the nucleus of an egg and hatch a dodo-like bird using one of the pigeons identified by Cooper's survey. It would, however, be almost impossible to recreate a perfect dodo, because its genetic code, which survives only in tiny fragments, could never be worked out to a sufficiently high degree of accuracy, said Dr Ken Joysey, a palaeontologist at Cambridge University. Source: The Sunday Times, 21 March 1999, by Steve Farrar.
There have also been concerns raised by ecologists that unextincting the dodo might cause people to believe that conservation isn't important since we can just bring back extinct species at will. But generally there's no reason why we shouldn't do it, and in fact it's already been done with the Quagga.
So it would be prudent if we preserved the dodo material that we do have more carefully for future generations. I'm not sure how DNA degredation can be prevented in such specimens, and it's possibly already being done, but historical precedent says that we always underestimate our future technological capabilities and motivation. For example, we let the dodo go extinct in the first place without anyone batting an eyelid; but now we'd rather like it back. And the same thing happened to the passenger pigeon, which was in fact possibly the world's most common bird not so long ago. Incidentally, a similar principle applies to archaeology: we're very good at recording sites now, but this wasn't the case just a century ago and we wince at the Victorian archaeological techniques now. Yet in a couple of centuries' time it could become possible to explore a site using a more advanced form of ground penetrating radar that enables us to learn more about sites than we do at the moment by digging them up and destroying them in the process. The general principle is that we must stop doing things that are irreversible, even if it means waiting a little longer to learn about things. We should either learn things properly, or not at all, otherwise we'll be left with incomplete understandings.
So at the moment, we've no idea what a dodo sounded like. If Andrew Kitchener is right (see e.g. the dodofact page), we're not even sure just how plump or otherwise it was. I really hope that if we've got a chance to find out, we're not inadvertantly squandering it.
Dodo Linkage
The dodo is most closely related to the Nicobar Pigeon which is a very pretty bird. It was possibly related to the Reuinion Solitaire, though sites seem to be a bit sketchy on this, and we have less information about the Reunion Solitaire than the Dodo, even though it was apparently more populous. Oxford have a page about the dodo with some contemporary sketches as well as their own relics and reproductions.
This follows on from a Swhack conversation, where jcowan has just mentioned that the quagga is not a good example of an unextincted animal.