There is a project called FatELF: http://icculus.org/fatelf/
LLVM IR wouldn't be a valid internal format, but only because it hasn't got an ELF machine type (AFAIK, anyway), and the loader would probably need to be patched to run it in the appropriate VM if there is no native version available, but that shouldn't be too hard.

Another useful, although far more difficult, project would be a VM which uses (a probably modified version of) LLVM IR, combined with a security wrapper around POSIX library calls, to produce an alternative to Mono/CLR, JVM, and so on, but usable with virtually all languages.

A state would be able to secede, but allow all US citizens to freely enter, settle in, and then vote in thier new country. However, as soon as they have done this, the US constitution no longer applies to them, so 5 minutes later they can change their constitution removing those parts.

I think the way to do it would be

1. amend the state's constitution so that it can re-define what it is to be a citizen of that state, but not actually change the definition in any way. This is legal because it doesn't change the privileges or immunities of citizens of the United States, but it cannot yet be applied in a way which would do so. The state may also need to amend the rules for amending its constitution.
2. secede, ensuring that all US laws are carried over (this is the hardest part), including the 14th amendment. This is legal, because secession is allowed provided the 14th amendment isn't violated.
3. amend the new country's constitution to remove the 14th amendment. This is allowed because any amendment can be removed with another amendment.
4. change the citizenship laws using the powers granted in step 1, which is now legal because the 14th amendment is gone.

The states had the right to secede (whether they still do is an interesting but irrelevant argument), and the Confederate states did so legally. The primary reason was not slavery, but protectionist laws which benefited the industrial north but which made farming (especially of cotton, but also other products) far less profitable (incidentally harming European interests, especially the British cotton industry).

The slavery issue was mainly raised to make Anglo-French military involvement more politically difficult (and it succeeded), but slaves in the north were not freed until later (the Emancipation Proclamation only applied to secessionist states).

Now, you can argue that Lincoln could declare war on the CSA as with any other war, and even that the war was a good thing, but using military force to prevent a state seceding was definitely anti-Constitutional, as was arresting the Maryland state legislature so they could not secede.

I think we might be talking at cross purposes here: AIUI, GGP suggested that a C++ to C translator could be written in a macro language, which would require a front-end (C++ -> C -> internal tree -> asm -> binary). For basic C++ this would be fairly straightforward, but doing things like template expansion robustly would be slightly tricky at best (although C++'s rules about space around the angle brackets would make this far easier to do than with, say, java's rules).

If I were trying to translate a C++ program to C, I would use one void* per thread (for the object being thrown) and then setjump at the start of the try block, and if it is the first execution, enter the try block. When an exception is thrown, we can assign the thrown object to the pointer, and longjump. The second time we arrive at the start of our try block, we are sent to the catch block.

If they used an early enough variant of C++, they would be relatively straightforward, since that has already been done once (cfront), but templates would probably make things somewhat harder, and would probably require a decent macro language. If they want some of teh nicer-sounding features of C++0a, things might start getting very horrible quite quickly.

I would assume they do, to reduce the problems anti-deeplinking scripts cause, especially for images where using them is more common..

I think he means you can turn up and study without enrolling for the first half, not that you don't have to turn up. Of course, a lot of places (I don't know about German universities) don't require you to turn up to lectures, or sometimes even tutorials, although not doing so is of course rather foolish.

This is pretty much equivalent to being able to drop our without penalty or record for the first half of the course.

No, the actual damage to the rights-holder is far less than the MSRP, because that's adding in the retailer's mark-up, the wholesaler's mark-up, and the rights-holder's marginal production costs (if the band holds the rights for music, that would probably be $0), not just the marginal profit and the proportion of the fixed costs which would have been paid by that copy. It might even be possible to argue that the actual damages were less, because you would not have bought it if you could not pirate it (I know that's true in some cases, but it would be hard (and probably pointless) to argue), which brings the actual damages down to quite a small amount per copy (for an MP3 rip of a song, the damages would be considerably less than $1 per song, since the iTunes price is obviously profitable). Then there is the fact that unless you had a very large collection, the damages would be less than the limit for the small claims court, so they would not be able to bury you in legal fees either.

This is the situation in Australia, (although Sen. Conroy has said he wants to change the law) and there are still plenty of record shops (my local mall has I think two record shops, and the nearest big one has 3 record shops and 2 DVD retailers, as well as 2 big retailers), so piracy can't be doing that much harm to the recording industry.

Yeah, that's a tactic which has worked very well for everyone who has tried it.

Python already has a DOM manipulation library (in fact, I think it has 2 or 3 which could be used in the normal distribution), the problem is browser support. If browsers don't use it, then web devs can't use it.

They could just append a LRO character to each post, which is a relatively small overhead, especially as it could be added as the page is generated, so they won't even have to store it.

Well, if they juts wanted to fix the sillyness which has been blamed for the introduction of unicode-mangling, they could just add a left-to-right override at the end of each post. That way, people would still be able to, for example, post arabic quotes, but they wouldn't be able to break anyone else's comments.

I suspect there are other "fun" things one can do with unicode, but they probably wouldn't be much harder to fix.

He didn't say they didn't come from commercial research, after all, anyone in a CS-related field would know of the contributions from Bell Labs and Xerox PARC. After all, there are probably millions of CRUD apps in use, but almost none of them can be described as particularly innovative, but the underlying technology (produced in either commercial or academic research labs, largely by people with PhDs), is. Certainly there are innovations which pop up in non-research software departments, or elsewhere (the Web, for example), but there is less innovation there.

I think his idea was that companies which use a steam-like UI to provide software in a manner which would have previously been considered piracy (possibly for a price, possibly with malware or just ads in the download system's gui). I suspect he's right, but people would still see the benefit of buying what we think of as legitimate copies (especially with the sort of marketing power organisations like the BSA have). How many do I don't know, but I suspect a large proportion of the home users who buy legitimate software do so because it is safer and because you get patches, not because they are afraid of being sued. Also, abolishing copyright wouldn't prevent companies using DRM to make it hard to use copies which did not come from them.

Businesses would be more affected by abolishing copyright, since they are generally interested in using many copies of the same software, but a site-licence could still be implemented using either a pre-sale contract, or by making use of the reason no one ever got fired for buying IBM: support and buck-passing. If you are using "pirate" copies, you lose that, and have to handle any DRM they chose to throw at you, and probably get no patches, so your system is likely to be very insecure after a short time.

You could probably replace most consumer EULAs with sale-time contracts, where the purchaser is required to sign a contract before they get the software. The contract could probably be somewhat similar to an NDA, except with usage restrictions on the software as well. The only difficulty I see is where software is bought by one person for another.