They've done worse, and let's not forget about the ongoing saga of the Celeron.

As of 2007, IBM's Blue Gene/P system cost $1.3M per rack, and the Blue Gene/L cost $800K (per a PCWorld story entitled http://www.pcworld.com/article/135334/ibm_drops_price_on_supercomputer.html). However, it should be noted that the hardware cost of such systems doesn't reflect the total configuration and operating cost. Many news outlets have reported on the favorable overall cost effectiveness of building supercomputing clusters with PS3s. Yellow Dog Linux has features specifically designed to support the Cell/B.E. CPU.

lmgtfy.com simply redirects the user to Google with the proper query in the URL, albeit in an amusing fashion. This is quite different from using a proxy system to query Google.

If you had bothered to pay closer attention to the article, you would know that you're wrong.

The article does not say Nikolaenko has been convicted of anything yet.

Slashdot titles are frequently fertile grounds for seeding jokes.

To my ears at least, you seem to be implying that Google is willfully complicit in something illicit. The act of searching for something does not break copyright law. Actually distributing it without the right to do so does break copyright law. Also, for purposes of increasing your chances of retaining some credibility in the future, please also note that the term "copyright" is not spelled "coppywrite."

You'll be waiting a long time. Google's AUP prohibits the activity you're describing, and it's extremely difficult to do it on anything resembling a large scale without getting blocked.

Why are you assuming I didn't review the study? I did, and again, the conclusions are deeply flawed. The appropriate course of action would be to instantiate improved policies for the production of documents that appear in PDF format for general consumption. Once again, the file format itself is not the problem.

Wow, the GP response occupies maybe 6x2 inches on my MBP display, and that's with the browser window occupying perhaps only 3/5 of the horizontal space available on the LCD. I think perhaps the issue lies with your particular parsing of the content.

Many applications can already export directly to PDF on exactly the terms you've described, and there are things like CutePDF that will allow you to "print" from any application to a PDF file with a couple of clicks under Windows. On Mac OS X and Linux platforms, you can typically just save any document as a PDF file, at least from most native apps. The capabilities you're describing are already in place, and there's no need to worry about strictly text and image-based docs you've created falling prey to any sort of vulnerability, at least not in the scope you've described.

For purposes of this particular discussion, I don't give a rip about security vulnerabilities from exploits in PDFs. We're talking about the accessibility of content that by all rights should be easily accessible to the vision impaired. On that note, it is firmly the fault of absolute morons (who are, in my experience, in the minority of content producers, by the way) who elect to convert text-based documents to PDF docs consisting of nothing more than one rendered image per page of content. It's actually beyond moronic, and I would formally reprimand anyone I found doing it with anything that was destined for even semi-public consumption. Above all else, it largely eliminates accessibility for everyone, including folks like myself who might idly attempt to search such PDFs for a specific term, and wind up screwed, because the idiot who created the document couldn't be bothered with operating an elementary piece of OCR software or gasp seeing if there was an alternate origin format that could likely be exported to a full-text-indexed PDF document with three clicks of the mouse.

The file format is not to blame. Morons who scan text-based documents into PDF files, saving each page as an image are to blame. Even in 1995 or so, when I was first exposed to OCR technology, it worked "fairly well." Anyone converting text to PDF by scanning pages in as images these days is a complete moron, and a huge variety of applications now support exporting text-based documents directly to PDF format with full text search and indexing capabilities intact, along with fancy formatting like gasp italics, bold script, superscript, subscript, numbers, fairly complex mathematical expressions, etc. Hell, images can even be embedded in PDF docs that are largely textual content (holy wow, the technology!), along with alternate text and hyperlinks. In other words, "WTFMATE."

There's a technical term for the practice you're describing: rubber hose cryptanalysis

Insightful?!? How did you manage to ignore the very real threat of asteroid impacts and other phenomena that are outside our control? It's really not a matter of "if" this will, but a matter of "when."

I've defined a contract in my user agent string, which is bound to wind up in their logs. It stipulates that every time I successfully load a page (HTTP 200) from their site, they owe me $100 USD. Should they decide to refuse payment, I have no reservations about issuing subpoenas for testimony from those who have administrative access to the logs and collecting what is rightfully mine. Let's hope for their sake that they're retaining their logs; I'd hate to have to have them brought up on charges of destroying evidence.