The whole issue of pseudonymity has come up again, both on Google+ and on ScienceBlogs. While I've been on the Internet for nigh on 20 years, my initial point of entry was through a Usenet group that strongly preferred real names (or something real-name-ish). As a result, I've never tried to maintain a separate Internet name-- all of my Usenet posting and all of my blogging has been under my real name. So I don't have a great deal invested in the question, on a personal level.
There are a couple of points, though, that I think are worth making about the recent discussion:
1) There's a much-linked story going around about a public health official who was ordered to stop blogging after an Internet troll complained. This is being passed around a lot with titles like the one Orac used, "The consequences of blogging under one's own name," but that's not quite right. What this shows isn't the peril of blogging under your real name, it shows the peril of blogging when your boss(es) would rather you didn't.
This is a known avocational hazard of blogging, but is entirely independent of the name under which you blog. If you have a blog, and talk about stuff related to your job, you are taking a risk unless your employer is ok with what you do.
Anonymity or pseudonymity is a way to duck the issue for a while, but no pseudonym is perfect, and anyone can be found out. What matters is whether your employer objects to what you do or not.
In fact, this lesson is clear from Orac's own post. As he says there, his pseudonym is regularly "cracked" by people who don't like what he says, who then proceed to write angry emails and letters to his employer. What allows him to keep blogging is not the fact that his real name doesn't appear on the blog, it's the fact that his employers don't object to his blogging, and are even supportive of his efforts.
It doesn't matter how careful you are about hiding your name-- in the end, you can be found out. What matters is whether your bosses will back you up when crazy people figure out where you work and start emailing them. If they approve of what you do, you're fine even if you blog under your own name (I have had people complain to my chair about stuff I said on the blog, to no effect, because he saw no problem with what I wrote). If they don't like what you do with the blog, it doesn't matter what name you attach to it, you're going to have problems.
Now, you can argue about whether employers can or should order employees to stop blogging, but that's ultimately between individual bloggers, their bosses, and the legal system. The important thing is that blogging about what you do for a living necessarily involves some risk, whether you try to hide your identity or not. It's your employer's opinion that is the key issue, not the way you sign your posts.
2) The argument against allowing pseudonymity is that people will behave better when their real name and reputation are on the line. The counter-argument is that this does not necessarily require real names, but rather consistent identities: someone who always blogs or comments under the same name will build up a reputation over time, and the reputation attached to that online identity can play the same role as the reputation attached to a real name.
There's a key qualification that's not usually stated, though: The reputation attached to a consistent online pseudonym can serve the same role as the reputation attached to a real name, provided the person cares about how that pseudonym is perceived. Which is not necessarily the case-- some people have consistent online pseudonyms which they use to engage in all sorts of jackassery, and either don't care, or make it a point of pride.
Does this justify a real-names-only policy? Not really. I suspect that most people who act like jackasses under consistent Internet pseudonyms would act like jackasses under their real names as well. Going to real names only would reduce the problem of bad behavior somewhat, but not eliminate it. But it's worth remembering that while consistent pseudonyms can function almost as well as real names in a reputational sense, that doesn't mean they will.
I suspect that most people who act like jackasses under consistent Internet pseudonyms would act like jackasses under their real names as well.
In fact, the very biggest jackass I know online does it all under his as-far-as-I-know legal, professionally-used, name.
I'll see Kate's biggest and raise her nine of the top ten.
IME, the worst behaved fall into two groups: Those who have nothing to lose and get off on being twits, and those who are true believers. The former have no reason to conceal their RL identity, while the latter think they are doing a good thing.
 Even if what they believe in seems to be in throwing a hissy fit that has been running the better part of a decade now.
Kate @1: You took the words right out of my mouth keyboard
Well, Mab*s used a consistent pseudonym, and everyone knew who he was anyway.
Females especially have added reasons and vulnerabilities (regardless of employer) for which they may choose to blog pseudonymously.
Well, I've stated in a few disc discussions on Scienceblogs that the network community has a history of real-name only policies.
First, real names were pretty much used by everybody on the Internet before the Eternal September (Google it).
Second, there was FIDONet (which was about 1 million users at its peak) which also had real-name-only policy (and what's more, the list of all nodes along with their phone numbers was public). I still remember FIDO quite fondly, the quality of discussions was excellent.
Spoken like a proper physicist. Pseudonymity still *helps* a lot.
Also, the definition of "bad behavior" is fundamentally a problem here. Obviously, you are talking about the definition of a privileged few.
Free speech isn't just a little thing in the American constitution, it's a good idea and a prerequisite for meaningful discussion on serious issues, and I'll fight for "bad behavior" any day when it consists only of saying the "wrong" words (minus really serious threats and maybe a select few other things).
Most users on the 1970s Arpanet used pseudonyms or their initials. Sometimes they'd use their real first name, but no other identifier. Of course, everybody knew everybody else back then, but I mainly remember people by their pseudonyms. I don't put my name on my door. I don't put my name on my car. I don't wear it on my shirt. Why should I need a higher standard on the internet?
I use a couple of different pseuds on different parts of the internet, and I think they're fairly transparent in the sense that anyone who pays much attention to what I say over time and wants to know who I am can figure it out. On the other hand, people (like hiring committees, say) who Google me won't find my voluminous internet comments. I think there's something to be said for that. The internet is not real life; things I say to friends or coworkers in person aren't likely to become easily accessible public knowledge, so I'm not sure why things I say to others on the internet should be.
(I think less-transparent pseudonymity is fine too, and I know some people with good reasons for not wanting to make it easy to figure out who they are.)
I ... I didn't originally post under my real name.
I was a different person then.
DON'T JUDGE ME! Seriously, senior year in high school, only person in town besides the librarian who even knew what the Internet was ... it was a different time, then. A darker age, when men knew how to use trn, and a killfile was a dangerous, feared thing. We fought September ... oh, how we fought.
There is a combination that is often ignored, that of employers knowing and permitting if the blogger remains anonymous. Then the employer has plausible deniability. It is not so much that they do not want the matter disclosed or discussed, they just do not want the burden of official admission.
Plausible deniability seems very useful, especially if the pseudonym never claims the authority of the job or institution. We keep talking about this as a legal rule, but I think of it as a heuristic for civility.
I remember lots of pseudonyms in the 1980s. I am grateful that the first system I was on had an eight-character limit on the names and I'm not trying to live up to "Wintermute".