"The television station also reported the company isn't incorporated in Rhode Island, but rather Delaware as a limited liability company, which would deem 38 Studios ineligible for tax credits in the state."

This is fraud if anyone but a famous, rich athlete does it.

While it's partially true that the industry standard pushed hard against the proprietary systems, that's no entirely true. Some of it was simply that the established proprietary system companies didn't try hard enough. It seems that color, video, and animation took Apple by surprise on the Mac.. they were offering what, the Apple ][ GS for that market? And they're the one that survived... only barely, and mostly because Steve Jobs came along and converted them into a high profit luxury CE company, not a personal computer company anymore (Macs are about 18% of Apple's business, and falling every quarter).

Commodore and the Amiga blazed that multimedia trail, with once revolutionary hardware and the best overall OS in the personal computer business at the time. But the bosses paid themselves more than the CEOs of Apple and IBM... combined, and spent way too little on R&D, despite Commodore at the time being more tied to custom chip development (eg, spending lots of money) than any of the others. Commodore couldn't remain competitive that way, but it was really more of a suicide than our being overrun by IBM compatibles.

Atari's management (the ex-Commodore Tramiels) didn't understand the difference between a late 80s/early 90s computer that needed real ongoing OS development at upgrades, and the Commodore 64 era of the OS basically being part of the hardware. Wang was already pretty much of it by then... they didn't really transition well from dedicated word processing gear to general purpose PCs. Rat Shack when totally IBM compatible, and eventually just didn't see any profit in making their own. Sinclair only put their toe into the 16-bit world anyway, with the very restricted QL, and hit enough trouble in 1985 to sell all their personal computer assets to Amstrad. DEC pretty much missed the idea that personal computers would grow more powerful than minicomputers faster than minis could keep up.

Water is by far the driving material resource. Metals are insignificant compared to water for human utilization. That Planetary Resources wants to track NEOs is also important. They have definitely done their homework.

When this news started to break last week it was unclear if they were just going after PGMs or had a more comprehensive strategy.

Electricity is the next step. Power beamed to provide propulsion and operating energy to other spacecraft. After that, beamed power to new space facilities, then gigawatts of green power to Earth.

Then water. Jim Head says "Follow the water." Find water sources that are at or near the top of the gravity well - Phobos, Deimos, Amor-Atens or Earth-crossing NEOs with high water content.

Those two resources enable reliable access to the Lunar and Martian surfaces. Minerals, lunar polar ice and many other elements are further down the list of needed (re: profitable) materials.

That said, such a well-heeled team has surely looked deeper into the specifics.

Vieques would be likely - Navy controlled territory in PR.

It happens... it shouldn't. But check out the notorious CADtrack XOR patent (US Patent 4,197,590). CADtrack was some kind of CAD company from way back, that emerged in the 80s as a patent troll. They had a patent on using the XOR operation to draw and undraw a cursor on a bitmapped screen. This should have completely failed the test of obviousness... I did exactly this as a kid of 17 on my first home computer, an Exidy Sorcerer (it didn't have full bitmapped graphics, but it had programmable character memory that could be used to simulate some game-class bitmapped displays with some cleverness). And at that point, I was mostly just self-taught "in the art", having read many issues of BYTE, Kilobaud, and Creative Computing. There should be a "kid rule"... if a High Schooler or below figured out your thing independently, you lose your patent.

A buddy of mine worked with the lawyers at Commodore on this, as we were being sued. He resolved to patent AND and OR... I don't think he could come up with the cash, though.

Sure looks like some monkey business here. This patent was filed in 2011, but it's actually a filed as a continuation. Here's the critical bit:

This patent application is a continuation of co-pending U.S. patent application Ser. No. 12/632,262, filed Dec. 7, 2009, which is a continuation of U.S. patent application Ser. No. 12/168,578, filed Jul. 7, 2008, and entitled DISPLAY SYSTEM HAVING FLOATING POINT RASTERIZATION AND FLOATING POINT FRAMEBUFFERING, which is a continuation of U.S. patent application Ser. No. 09/614,363, now U.S. Pat. No. 7,518,615, filed Jul. 12, 2000, and entitled, DISPLAY SYSTEM HAVING FLOATING POINT RASTERIZATION AND FLOATING POINT FRAMEBUFFERING, which is a continuation of U.S. patent application Ser. No. 09/098,041, now U.S. Pat. No. 6,650,327, filed Jun. 16, 1998, and entitled, DISPLAY SYSTEM HAVING FLOATING POINT RASTERIZATION AND FLOATING POINT FRAMEBUFFERING, the disclosures of which are incorporated herein, in their entireties, by reference.

In short, they get to claim a priority date going back to 1998, as long as at least one patent application in the chain was in pending status when the next was filed, going all the way back to 1998. And that at least one inventor is listed in common. And, of course, that this filing (8144158 is all about the format of their floating point) is really an elaboration on what they did back in 1998, and not something different.

This is a really long time for this kind of thing... and that's just what they were after, of course. At least, under the new patent laws, the filing date and 20 year life are based on the actual filing date. Under the old system, you were given 17 years from the date of grant, so it was common to try to delay the grant date as long as possible, so that the things you did "way back when" were now commonplace. A legal submarine patent. This long string of continuations amounts to much the same kind of thing, though of course, they can't claim something different. For example, the 32-bit or 64-bit floating point units in current GPUs are probably not infringing, unless they derive directly from this design (unless it's determined that the FPU design is in itself obvious, or already covered by prior art). This also means that the 8144158 patent will expire in 2018 along with 6650327, not in 2031, which would be the expiration date of a new patent applied for in 2011.

From the troll company's viewpoint, it could well be that they found a bunch of GPUs out in the market using a design very similar to that of the old SGI GPUs, and decided to file this to enhance their case, which again, is just dandy as long as at least one patent in that long chain was still pending at the time this one was filed. It could be the defeat of the '327 patent against ATi illustrated their weakness going up against other devices.

I wondered who they're specifically after:

Apple iPhone: PowerVR (all. SGX543MP2 in the iPhone 4S)
HTC EVO4G: Qualcomm Adreno 200
LG Thrill: PowerVR SGX540
RIM Torch: Qualcomm Adreno 205
Samsung Galaxy S: PowerVR SGX540
Samsung Galaxy SII: ARM MALI-400
Sony Xperia Play: Qualcomm Adreno 205

So that's basically every GPU maker in a mobile device other than nVidia. And not all that interesting that Qualcomm's on the list, since they're now licensing PowerVR anyway, and so Adreno may be relegated to economy devices, or gone completely in short time. This was probably critical to Qualcomm's reputation, with the Krait-based S4 SOCs, they'll otherwise have the fastest per-CPU-core ARM on the market, at least until A15s are out. No need to hamstring that with the weakest GPU... but I digress.

The patent clearly can't cover all uses of floating point, and even the patent authors clearly understood this. But claims, as always, try to suggest they cover everything.

In truth, a patent covers a specific way of doing something. So assuming there's no other reason to judge this a bad patent, you still have to look at exactly how they implemented their floating point solution in hardware. For example, some fragment of SGI went after ATi some years back, and the court rejected the '327 patent as not applying, even though ATi was also using floating point rasterization. If your FPU engine uses a similar floating point implementation and architecture, even if it's an improvement, it probably infringes. If you went an entirely different way but still happen to use floating point, probably not.

Obviously, there are edge cases, which is the whole point of having a patent court -- it's not always obvious if one device infringes on anothers' patent. And the courts aren't a panacea, either, they sometimes make the wrong technical decisions. Just the best system that exists for this kind of thing so far (though in terms of digging up prior art, the community as a whole often does this public service much better than the courts). Keep in mind that ultimately, the claims themselves don't have legal weight, other than to illustrate the part of the patent the patentee claims is unique. You have to look at the text of the patent when deciding what infringes, and specially, the very specific preferred and alternate embodiments called out in the main description. Many have tried, but in theory, you can't patent an algorithm.

That's incorrect -- the description is the basis for granting the patent, it's the part that actually matters. The claim matter, too, but in a different way. They're calling out parts of the invention description that the inventor claims are unique. But the patent is granted on the basis of the description, not the claims. The claims are checked, but not with any precision. It's quite common for claims to be judged overly broad, or thrown out entirely. They have to relate directly back to the description of the invention, and they act largely as footnotes -- they can suggest, for example, that deviations from the preferred embodiment of the invention (what's discussed in the description, for any utility patent) are just dandy, but that in itself has no legal power. The PTO will usually eliminate overly broad claims, but not always. It's the body they're primarily concerned with -- that IS the invention.

The claims are for the courts, and very much left open to interpretation. In fact, when you write a patent, you KNOW that the PTO is not overly careful in reviewing claims, and so you try to make the claim sound like it's covering anything you can get away with. But, in fact, it doesn't. A patent covers a very specific thing, not a general case of all similar things. This is one reason that companies, particularly patent trolls like this fragment of the once great SGI, don't like to go to court. Once in court, a claim can be judged to not apply. But it can also be tossed out entirely, either based on being overly broad, or in light of prior art. When that happens, you can't use it again, even to scare someone else into licensing. If significant enough, it may even call the PTO to re-examine the entire patent.

Actually, the claims only have a certain level of legal weight. The actual patent is granted on the "Description of Invention" -- the main body of the patent. That's the part the PTO reads and analyzes against prior art. The claims are generally only checked against the body, to ensure they're supported by the body, and I'm sure, not written too outlandishly.

Now, once you're in court, it's the claims that are used to prosecute infringement. But they'll be challenged, or even tossed out, if they aren't properly backed up by the body of the patent, once you get to trial. It's true that you can't claim infringement other than on claims, but the claims are not the final answer. They're more like an annotated index into your description of what you believe the actual "invention" to be.

I'm not a lawyer. I do write patents, and analyze the technical merits for lawyers, from time to time. Not my finest work, but I'm good at it.