Jimmy-
What would be different is how users are dealt with during the intermediate period between the trouble starting and final adjudication. If a quickpoll indicates a ban, the user is banned unless and until the appeal is successful.
Well, I invented quickpolls so I'm of course in favor of using them, but I must caution that results so far haven't made me very optimistic. There have been a couple of cases where there were quickpolls on two users for violation of the three revert rule, and one of them got banned, and the other didn't. It was clear that the user who didn't get banned did so solely because people liked him more ("we can't afford to lose him") or agreed with him politically.
So this does to some extent confirm the fears that people have had that voting would lead to popularity contests instead of informed decisions. It may be possible to address this through a "same crime, same time" policy, moderation, restrictions on voting, and so forth, but with the current system Fred is correct when he says that quickpolls are good for flies but not for hawks. That is why we have presently agreed to suspend the use of quickpolls for violations of the three revert rules, and use them only for more obvious cases. Policy reform is of course welcome.
We also need a wide range of instruments of punishment:
Technical: - per article bans - edit throttling ("edit only every x minutes") - edit limiting ("only edit this page x times per day") - namespaces limiting ("no editing of Wikipedia: pages") - deny access to page history (makes reverting difficult)
Policy: - "no revert without prior discussion on talk page, otherwise ban at discretion of sysop" - "no reverts, period" - "remove personal attack X, or else"
Regards,
Erik