Hmmmm... well you know SOME people are going to be doing it, and if it works they will have one up on everyone else. This is a really tough issue.
On one side, one could say since the average Seti@home Windows user will just upgrade and let her run as the good old screensaver, the bandwidth used from them will greatly decrease, giving hardcore Seti@home teams space to keep pumping in the WUs.
On another side, if all the teams get "greedy" and run the 3.0 client thru shserver2, it may defeat Seti@home's work to decrease the bandwidth they use with the 3.03 client. This sort of falls in the realm of car radar detectors, overclocking, things like that. Can't come up with a perfect analogy at the moment though...
On another side, we don't know if getting 3.0 or older results will mess up the back-end processing or not. From the article that someone posted last month, Seti@home is getting ready to do some major back-end processing, which one would assume that there will be a LOT of old client-processed WUs getting worked on. I didn't realize v1.01 was still good, but my friend just had to upgrade from 1.01 to 3.03 a few days ago according to the message he got from Seti@home, so apparently they were even accepting 1.01 client WUs as well.
Will running 3.0 screw up their back-end processing? IMO, I doubt it, but it is possible. Will it defeat the purpose of limiting incoming data bandwidth? I don't know. Whoever talks on newsgroups or wherever with the Berkeley Seti@home folks, please bring this up. We KNOW people are going do it, we just need to learn if it is "OK" to do it or not. If it is "OK," then we SHOULD all do it, I think...