Is web-presence important for researchers?

The short answer, at least in theoretical computer science, is yes. Especially pre-tenure.

The Coin of the Realm in academia is fame. Hiring and promotion decisions are based primarily on the perceptions of your impact by leaders in the research community. Those intellectual leaders must know who you are, they must know what you do, and they must think that what you do is excellent. This is precisely why it's so important to network, network, network — go to conferences, visit other departments, talk to visitors, ask questions, answer questions, go to lunch, drink beer, play pool/golf/frisbee/Settlers of Catan, race go-karts, exchange business cards, all that stuff. Having a visible online presence is just another form of networking.

Similarly, if you want to attract good students, they have to know who you are, they have to know what you do, and they have to think what you do is interesting.

Similarly, if your work is not freely and easily accessible on the web, it is much less likely to be cited than freely accessible work of comparable quality.

To give some personal examples, I have good reason to believe that these web pages were a significant factor in my academic job search and even my tenure case, and this stuff definitely helped me get promoted. I expect that these pages similarly helped Suresh, and these pages similarly helped David.

In my opinion having an up-to-date website is very important. This way people will find out about your papers and what you are working on. It can also help to attract students. It also helps when you're teaching as you can put frequently requested information there and save yourself the hassle of having to reply to hundreds of emails.

In my experience, maintaing a website is not that much effort either.

I think it also depends on the position of the researcher. For instance, a computer scientist who has a permanent research position, publishes one paper a year in a top conference, and doesn't care about the publicity does not need to have an important web presence.

On the other hand, a postdoc who is looking for a job knows that every time he sends his CV somewhere, one of the first reflex from the recruiter is to Google his name. So, in this case, it's quite important to have a good presence, and to have an updated webpage, in particular with papers accepted but not published yet.

About the maintenance, as mentioned the other answers, maintaining a website is not particularly demanding, especially if it's quite simple. However, maintaining a blog can be quite complicated, especially because having an non updated blog is probably worse than not having any blog at all. As for G+/SE, I guess the investment is worth the return from the community.

Concerning the danger of a high web-presence, well, obviously there is the risk that some "private" information might be connected to a public profile. For instance, I have a flickr account with pictures that, although not particularly shameful, I wouldn't like a potential employer or a student to see. Of course, my account is under a pseudonym, but that's the same pseudonym that I can use on other services (such as twitter), and maybe at some point I will refer to my twitter account from my G+ account, that maybe I will refer from my SE account, where I use my real name. But I guess that's the risk with the Internet in general: if it's out there, it has to be considered as public. There is a similar argument for opinions or ideas you could have a site such as Academia SE and that could be later on taken out of context and used against you.