A referendum on switching to the Alternative Vote (AV) system will be held - assuming that the coalition holds together long enough to pass the legislation and that the Lords vote for it - on 5 May. The details remain to be seen, but I will unequivocally be supporting a "Yes" vote for a change to AV.
Firstly, while Alternative Vote is not a Proportional Representation (PR) system, where the political views of the MPs match the votes of the country, it is still a significant improvement on First Past the Post (FPTP).
I would much prefer a PR system such as Single Transferable Vote (STV) or open-list PR. However, that's not what is being offered, it is not going to be offered - the Conservatives would simply not vote for any Parliamentary amendment to that effect - and it would be extremely unwise to include it in the referendum anyway due to the risk of splitting the pro-reform vote.
Alternative Vote is, however, one of the best constituency-level systems. If you are going to keep single-MP constituencies, and that's not likely to change for a while, it allows people to vote their true preferences, avoids the majority of the need for tactical voting, and reduces the number of 'wasted' votes.
Voting against an improvement on the current system - even if it doesn't go as far as I'd like - seems absurd. If the referendum is won by the "No" vote, then the political classes will not say "Maybe the voters are telling us this was unacceptably small - let's discuss PR". They'll say "There is clearly no public demand for voting reform" and we are unlikely to get another opportunity in the next few decades.
AV would get people (in England) used to preferential voting systems, as well, which would make a later step to STV or the suggested AV+ much easier.
Secondly, the system would significantly reduce the number of "safe" seats, making it easier for unpopular or corrupt MPs to be removed from office. There are seats where the incumbent MP has over 50% of the vote anyway, which would be unlikely to change hands under Alternative Vote, but even they would be vulnerable if the MP did something particularly wrong, as there would be no risk of splitting the vote against them.
Because of this, it would be important not just to get as many first preferences as possible, but to ensure that your party is an attractive target for second preferences. Certain forms of negative campaigning that detract from the public opinion of politics but nevertheless are effective would cease to work. Parties would have to be able to pick up those second preferences, which means not being too negative towards the first preference party of those voters.
Rather than attacking other parties on the grounds that they wouldn't be any good, there would instead be a requirement to convince voters that your party's policies would be better at solving a particular problem, or at least almost as good.
The electoral experiences of Papua New Guinea are a strong indication that this would change how campaigning and political positioning occurred, and for the better.
This, incidentally, is one reason that I don't hold too much faith in "what would have happened under AV" calculations for the recent general elections (though they are worth doing to get an idea of the minimum effect). You can calculate the effects of transferring votes easily enough, but that doesn't take into account that the second preferences - and perhaps the first preferences - would be different in a different electoral system.
People often use these figures - pointing, for instance, at the 1997 Labour landslide and saying that AV can be less proportional than FPTP. Firstly, campaigning would probably have been a bit different under AV. Secondly, if the 1992 election had also been held under AV, it would probably have resulted in a Lab-Lib coalition. Kinnock doesn't resign, and the 1996 general election would have been contested on an entirely different basis.
Thirdly, a move to AV for general elections would probably also cause local elections to be run under AV (or perhaps even STV for the multi-member wards), and almost certainly also replace the bizarre Supplemental Vote system used for electing mayors. This would go some way towards reducing the number of different electoral systems in use in the UK, and simplify things for voters.