I'm going to start using SVK for work, but I'll continue to be interfacing with our main repository which is CVS. Any tips and tricks people can offer me for making the two work together would be appreciated. If I understand correctly, SVK can't currently just talk to CVS as its backend like it does to Subversion, so I'll basically have to manually commit to CVS, and when I do that people won't see the intermediate changes I've made to my own local SVK depot. That's probably desirable behavior at this point, although I'd rather SVK manage talking to CVS for me.
If someone can persuade me that Git or something else talks to CVS and would be better for me, please do so. In terms of experience I feel pretty competent with CVS, moderately competent with Subversion (though I've never been comfortable with the way tags and branches were flattened into the filesystem namespace, and I never really learned how to weave around between branches; I figure I'll just skip learning that and depend on SVK to do it for me under the hood), and I have a week of experience with SVK. I watched part of Torvalds' rant where he described git and understood what I saw, for the most part.