Ummm-- nonsense? Lots of people have one person, above all others, that they cherish. And for most of life, it doesn't really matter that this is the case (but like, for most of life, my hobby doesn't actually much matter to other people). But sometimes, that one person I cherish really does have extra power. Who gets to visit me in the hospital (answer: the one I cherish does!) Who gets my kids if I die (answer: the one I cherish does) Who gets my stuff if I don't have a will? (answer: the one I cherish does).

And hey, isn't it handy that there's a super-simple, cheap way to tell who I cherish: it's the person holding the marriage certificate! So an entire mass of horrible, messy, expensive problems becomes simple and clear.

Oh, and it also turns out that there's a nasty problem with the way that humans procreate: it's really long term, only one gender can do a bunch of the hard work. And often people who cherish each other have a commitment that one will do more of the looking after kids and the other more of the earning money. And because it's two people that work like one unit, it makes sense to fiddle the tax codes a bit so that it's more or less fair. (Like everything in tax code, there's always corner cases)

WTF? We're talking about bulk here; the normal 95% of people in a profession, not the nobel prize winners. I've literally no idea how you managed even think that "rockstar" was the key issue here in any way, shape or form.

You're not trying very hard to find any counter-evidence, are you? The fact that other STEM fields are experiencing increasing balance, and our is increasingly unbalanced doesn't register for you? The many personal anecdotes are not in your site?

Worse, you don't see the increasing evidence that men and women are much, much more alike than non-alike? That both sides are fully capable of essentially all tasks that the other can do?

In ever so many fields, over the last hundred years, men have declared that only men can do job A, B, or C; it's been clearly proven wrong in basically all cases. Is our field so very different? History would say no: we are like cooking (once a male prerogative), telephone operator, surgeon, and CEO.

Damn straight. IBM had a network, and Burroughs had a network, and ICL had a network, ... everyone had a network! And nobody's network talked to anybody elses.
Except IBM. They have about four incompatible network schemes :-)

Yes: they would have been one of "n" winners, each with incompatible content. You'd be in the situation (like the old phone companies) where a person on network "a" couldn't contact a person on network "b". That would be substantially less valuable than the fully interoperable internet we have today.

You're rather cherry-picking your data. Solyndra made a big bet: that the raw cost of the silicon in solar power would be important, and that a remarkably cool manufacturing technique to use a lot less would have a ton of value. As it turns out, that's not how the industry went: silicon costs dropped faster than anticipated, and the manufacturing costs of the Solyndra didn't.

We weren't "picking winners and losers" here: we enabled a big bet. Big bets don't always work.

And the internet was absolutely funded for years by the public purse to develop all of the major technologies and to make the same set of "big bets" about the valuable and non-valuable aspects of internet communication. Private people only became interested because of that investment.

And part of the investment was the "picking a winner". The key to the internet is that it worked across multiple vendors. If we hadn't have done that, there would be an ATT network, an IBM network, a Unisys network, and so on. The government chose a winner (cross platform) and a loser (per-company networks).

Yes, but that not how IP networks work. When the server sends you a packet, it needs to pick exactly one IP address as the destination. Because your WiFi and Cell are on different networks, they give you different IP addresses. So the server has to pick either your WiFi or your cell IP address. Once that packet is sent, it's not going to ever get to you via the "other" network.

That's why the multipath needs special support. Among other things, lots of web sites which are on multiple load-balanced servers need to affinitize your session to a single server. Those load balancers are currently (AFAICT) knowledgeable about Multipath.

My prediction: apps will have to opt-in to get this feature, but beyond setting a flag when they set up the connection, nothing more is needed.

New protocols? A low-level protocol like like a PGM or ICMP? No, the RT sockets don't let you do that (among other things, there's hardly any value: even if you made a new low--level protocol, you'd have trouble getting internet-scale adoption (heck, even useful things like PGM have trouble, and we're never getting another ICMP again).

RT Sockets are a wrapper over WinSock (aka, Windows version of BSD sockets), but with some stuff cut out and object-orient-ified.

Links: documentation is at http://msdn.microsoft.com/en-us/library/windows/apps/windows.networking.sockets.aspx/
and there's a talk: http://channel9.msdn.com/Events/Build/BUILD2011/PLAT-580T/