Or they could have included three or four more copies of the lander and still cost less than sending humans. Rosetta has been in space for 10 years-- there aren't going to be humans floating around in tin cans in deep space for that long for a *long* time. At least not live ones.

That's not an argument for manned missions so much as an argument to either make things we're sending smaller and more capable or increase our ability to send larger and larger things. A significant portion of the mass you send on a human mission will be just stuff to keep the people alive, limiting the amount of stuff you can send for them to do experiments or massively increasing the cost because you're trying to send a general lab. If you're developing a set of very general lab stuff that humans can take halfway across the solar system, for the foreseeable future (even with significant reductions in cost to orbit) it's going to be less expensive and lower risk to make a robot that operates the lab and feeds it with stuff.

Delivery vans often reduce other traffic-- a UPS truck going through a neighborhood dropping things off might be eliminating a bunch of local trips to run errands, reducing the traffic on local roads and the need for parking at local sellers.

When I was using a bike for primary transportation and we got 3 feet of snow in one day (and another foot the next) the only reliable transportation was bike. I helped push a lot of stuck cars out, only to watch them creep along and get stuck again.

Bike gets stuck? Pick it up and move it. Get to a street that's not plowed? Carry the bike. I went grocery shopping by bike in that mess, and it was *way* better than any other transportation.

Or Herschel & Planck?

http://orbitalhub.com/?p=475

They don't stay together, but they're large satellites launched in a stack. The Herschel load doesn't go through the Planck structure, so maybe it doesn't count.

Word actually has a "comment" feature that attaches comments to particular locations in the text, and a flag that lets you show/hide them. It's been there for quite a long time. It works quite well.

Yes, they track usage. They know what page you're on.

3% to 5% failure rate is about typical for more reliable launch vehicles.

It's not like airplane flight yet.

The NASA centers that buy spacecraft and instruments still do some in-house builds, partly to maintain the skill base so they can be smart buyers. For instruments they tend to build things that are first of a kind. But even in those cases they're part system or instrument integrator, part builder, because nearly all the pieces are purchased from outside shops.

It also rewards lying, cheating, and fraud, which have all increased in science as the competitive pressure (for positions and funding) to publish large volumes of high-impact research in short times has increased.

Overall an excellent post-- too bad you posted anon so fewer will see it (I don't have mod points).

To expand on the quoted section-- good management of technical people is finding the areas where people are exceptional and putting them into positions where you can best use those skills, and avoiding forcing them into areas where they aren't particularly special or interested interested in becoming so (which will make them unexceptional).

I should also point out that there's a very fuzzy line between "engineer", "scientist", and "inventor" (which you sort of imply, but I'll come out and say it). A very substantial part of many science fields is engineering to make the observation that you're trying to make. And anybody can "invent" something that they need, but it usually takes a bunch of engineering to implement it, much of which gets done by the inventor. Even on a small scale, a significant fraction of engineers have to "invent" things on a small scale on a regular basis.

Or it could be because they're a disproportionate fraction of the cars on the road in California even if most people who drive them are pokey, when you multiply the numbers by the rate, you still get into the top 20. I've seen Toyota lots in LA that look like they're the storage lot at the end of the Prius assembly line-- hundreds of them as far as you can see. On the road, it's pretty rare to not be able to see at least one in your field of view. The C probably made it onto the list because it's more likely to be driven by younger people who aren't hypermiling yet.

But nothing says *all* your power has to come from the grid. It's entirely possible to switch some circuits of your house off grid with a second panel, or even a single switch that lets the house be connected to the grid or the personal system. The house is still connected to the grid, you just don't pull much off it.

I have a peculiar case of unreliable power from the power company, but not enough roof to make a full grid tie system worth the trouble (trees over the house block most of the sun). I'll eventually get around to putting in about a kW of solar and some batteries that will power my outdoor stuff normally and that is usable for minimal power during outages just by running cables. It's not much more than the cost of a good generator and doesn't require going to get fuel (which could be difficult after an earthquake). A friend of mine has a grid-tie system that switches completely off the grid over to batteries when power goes out.

And never worked large projects.

For some projects (not necessarily software, and not even necessarily that large) there aren't more than 2 or 3 companies that can do them, and they're all a pain to deal with.

This is a part of why the government created Federally Funded Research and Development Centers (FFRDCs). DOE created them during WWII because they saw a need for an organization that had more flexibility than the government (they're all privately operated) but would act as an expert exclusively on behalf of the gov't. They get their money entirely from a single agency and most do a combination of direct work ("we'll do it in house") to main expertise and procurement ("we'll spread it around"), recognizing that to be good buyers they need to have a good deal of internal expertise. Most of the FFRDCs are run by DOE (they started as the weapons labs and science centers to develop knowledge of elementary physics for weapons research), but many agencies have them. It's probably time to create one for government software development, or assign is as a role to a few existing ones.

As far as I can tell, Sheldon is a string theorist and is doomed to being a perpetual post-doc. The others will all have much more successful careers.