We have no idea what portion of the universe is observable to us. It could be a small fraction, or a large fraction, or even more than 100% (we could be seeing parts from different directions). If we can only observe a small fraction, then it's possible that the universe is half matter and half anti-matter, with the boundary beyond our ability to see.

However, science only deals with what we can observe, so there's not much point in worrying about the possibility. We also assume that there's nothing particularly special about our art of the universe as compared to the rest. That's served us pretty well, but it could be overturned by observation -- for example, our location within the galaxy is actually a bit atypical.

Science fiction is not about science.

A very good example of this is the Hitchhiker's Guide to the Galaxy. I love this series -- it's one of my all-time favorites. But it's not about science, and doesn't even involve much real science. It's a parody on modern society. Most science fiction is more about modern times than anything.

If you want them to do book reports, then have them read real non-fiction science books. There are plenty out there, quite a few suggested in these comments.

But if you really want to inspire kids, I could not recommend anything more awesome than the Open Research thread above by jbeaupre.

I've seen many good explanations for why estimating software is so difficult. But my favorite explanation points out the fact that when we write software, most of what we write is not like anything we've written before.

Engineers in other fields don't run into this as much. Building a new house is very much like building any other house, so they've got a pretty good idea of how long it takes to build a house like that. And even then, building houses typically takes longer and costs more than the estimates.

I'm not convinced that we'll ever have facial recognition software that will be able to identify anyone in the USA. The false positive rate would have to be below 1 in 300 million to be completely automated. That's a really high bar to achieve. In addition to your points about poor quality images and various angles, there's also the fact that people's faces change as they age. Other things can fool recognition software too -- facial hair changes, facial expressions, makeup.

My guess is that something else will come along that will do a better job than facial recognition before it is "perfect". Perhaps cameras so good that they can see fingerprints or teeth. Or heartbeat signatures. Most likely a combination of several biometric sensors that will monitor public areas.

Facebook's recognition software doesn't look through its millions of users to find a facial recognition match in a photograph. It only looks through your list of friends. So it's only going through a few thousand photographs of a few hundred people. It doesn't have to go through millions of photographs.

Of course there is. There's plenty of commercial demand for gold outside of its use as money. Jewelry and electronic circuits come to mind right away. Wikipedia lists quite a few other uses.

That sounds just right. Roswell was in 1947, and it would have taken a few years to reverse engineer it.

Seriously, I wanted to mod you up, but just lost my mod points. You make the point very concisely.

But honestly, I think you're actually understating the case. It's very likely that 1951 is a lot closer to the actual invention of the integrated circuit (1957) than this paper is likely to be on the origin of life.

So they're going to be bigger than Google's data center in 3 years? That's interesting, because nobody really knows how big Google's data centers are. I'd be willing to bet that most of the server admins at Google have no idea how many servers they have. The current estimates are approaching 2 million servers though. I have a very hard time believing that Iran is going to be able to build 2 million servers and the required infrastructure to run 2 million servers in the next 3 years.

I'm also curious why they'd need that many servers to run a mapping service. Their service would pretty much be guaranteed to have fewer users than Google, and provide only a fraction of the services that Google does. So I guess they're admitting that they don't really know anything about utilizing servers effectively. Or perhaps we should assume the more likely scenario -- that they're completely making all of this up.

So a tax-law professor thinks that his taxes are paying for Google lunches? If he's this confused about simple logic, how could we expect him to have any understanding of complex tax laws? He's obviously too incompetent to do his job.

For TextMate and TextWrangler, take a look at Sublime Text. A lot of TextMate users have made the switch, and it has the advantage of running on Mac, Windows, and Linux.

The Green Revolution preceded Earl Butz by a bit, and had a similar impact.

This doesn't sound at all ground-breaking. They've basically discovered what Hadoop already does -- if you shard your data, it makes sense to run the processing where the data is, to reduce communication overhead. And Hadoop didn't pioneer the idea, either. It's based on Google's MapReduce, and I'm pretty certain that the ideas go back much further than that.

Because Microsoft makes more money on a sale of Office than it does on a sale of Windows.

Did you not see the Slashdot article/video yesterday, showing off the Solowheel? Powered gyroscopic unicycles are pretty darn inventive/innovative. As was the Segway.