Is it ethical to withdraw a CS conference paper after being accepted, but before the final version is due (because of authorship disputes). What other consequences would there be (like re-submitting to other conferences if withdrawal is is successful)?
If you are able to resolve the authorship dispute, then clearly that would be the best solution. On the other hand, if you are unable to resolve the dispute (at least on a timeline compatible with the conference's schedule), then not only is it ethical to withdraw the paper, but in fact that would be the only ethical course of action. Basically in this situation you have no reasonable choice and you simply have to withdraw the paper, which is why it is clearly ethical despite the fact that it means you will have wasted the time and effort of the organizing committee and reviewers.
With that said, even though withdrawing the paper is ethical, there may still be some hurt feelings and a small amount of loss of credibility on your and the other coauthors' part, so it is important to do all you can to minimize the damage by being as honest and transparent as you can about what happened and about your motivations. As Anonymous Mathematician points out, just telling the organizing committee that there's an authorship dispute might come across as a lame and possibly suspicious explanation. What I would do is make sure to include in my withdrawal email:
- an apology;
- a relatively detailed explanation of the nature of the dispute, which establishes the claim that withdrawing the paper is the only reasonable course of action open to you (if there are some highly sensitive or personal details, you can omit them, but try to provide as many details as possible to make your claim that you have to withdraw as credible as possible);
- an apology! Specifically, acknowledge that you should have sorted out the authorship issue before submitting the paper and that you and the other authors are at fault for not taking proper care on this matter. Make it clear that you understand what went wrong and will be more careful in the future.
My feeling is that with the proper explanation, no one will bear you any hard feelings for creating this somewhat awkward situation -- reasonable people understand that these things can happen (as Anonymous Mathematician points out, in rare cases they can happen even when everyone is behaving reasonably and has the best of intentions). Your reputation will survive. In any case, letting the paper be published when there are unresolved issues surrounding authorship is almost certain to lead to a much bigger mess and much greater damage to the reputations of everyone involved.
You should be extremely careful with this. It's possible in principle to have an authorship dispute in which everybody is behaving completely reasonably and ethically (and it's just a really delicate, borderline case), but generally this means someone is behaving badly. If you announce to the organizing committee that there's an authorship dispute without clarifying, it could give everyone involved a bad reputation.
It's also unclear how withdrawing will help you (since that won't resolve the dispute), unless there's a newly added author who just doesn't want to publish in this venue.
IMHO, withdrawing a paper is never a good practice, and this is much worse if you withdraw after acceptance. Indeed this means that you (and the co-authors) asked for a peer-review, several reviewer worked (for free) on your paper, then a committee accepted it, and at the end you decide to discard all this work.
Moreover, reviewers give you also some useful advices. If you decide to withdraw, then one can imagine that you will use those comments for improving the paper and (for instance) submit it to a journal or to another conference. And this would not be fair...