PDFS--Of the WotC Court Case

[f33b]

Thanks for the docs, interesting read. Does section V1, 1 ("Prayer for Relief") have anything to do with the PDF taken down? Is DriveThruRPG/OBS a person "in active concert or participation" with the named defendants?

Also, regarding the pixel wide watermark, I'm not an expert, but I do work in a company that produces print and ebook quality PDFs (honestly, I work in a different department), but running some kind of diff against two or more pdfs isn't as not trivial as some posters make it sound. Remember, PDFs arose out of Postscript languange, these aren't flat image files like jpgs or gifs. Much of the content is encapsulated as vector graphics. Even if the micro water mark is an embedded raster image, you would probably have to parse the underlying Postscript code to identify the stream containing the image data. Removing this data will, in all probability, break the PDF (make it unreadable to readers, corrupt the data.)

edit: to clarify, (or not) it is conceivable that the micro watermark is a one pixel wide, transparent image object that is present in the pdf as a transparency, and is placed over or behind a larger image on the page.

You could harvest the text out the PDF, but would be left with a flat text file, and most likely would have issues with text being lost due to unsupported fonts and crazy formatting to boot.

If we look at the complaints, only three defendants were actually identified as buying the PDFs, the others, including the three John Does, "just" redistributed the watermarked PDFs (obviously, the meat of the complaint).

I think WoTC showed their hand too soon on this one, better if they had waited and identified more individuals involved with the initial purchase and infringing event, so to cast a bigger net. Of course, letting it be known that there is a watermark which 99.99% of all users will not be able to see, find or do anything about should do something to curb (but not suppress, people do have scanners after all) pirating.

WoTC could also look into using something like Microsoft Reader, which is a pretty heavily DRM'd eBook platform, but I doubt that would be widely popular, as the most secure form of Microsoft Reader requires users to have a Microsoft Passport account.

 

[fanboy2000]

It doesn't matter how much the court grants in damages. Real damages that can be collected can only go as high as the defendant can actually pay. The courts can award millions in damages, but WotC will be lucky to collect tens of thousands.

Very true, but a high damage award could mean a line on the defendant's property and other un-fun things I'm sure he never in a million years thought he was going to face.

 

[tomBitonti]

My (vague) understanding is that Watermarking technology is (or has to be) much more robust than the simple strategies which are suggested. Aren't there ways to hide the watermark in what appears effectively to be noise added to the data? (And which shows uncorrelated and random differences between any pair of documents?) But there are very bright and hard working folks who make their entire careers on this sort of stuff. I have to figure that there is an amazing amount of sophistication in the technology.

 

[Silverblade The Ench]

Question, since I didn't read the court documents:
How do they know the accused actually put these items up on P2P themselves?
How can they prove that someone didn't hack their PC and take the files, or a friend using that PC, considering how many folks' WIFI networks aren't secured, PCs that are shared, and how many folk's computers are riddled with malware?
Or that the security system has itself been broken..rather neat way to cover your tracks by putting Santa Claus or some poor schlub as the actual ID, eh?

I know civil cases require less proof than criminal, but the burden of proof etc...ya know?

And the concept that copyright cases can hit ordinary Joe Bloes harder than criminal fines...is outrageous.

And for the record, I support D&D by buying it...rather a bloody lot of it over the years. My concerns are about this copyright and IP madness.

 

[HeinorNY]

What's the point in sueing a guy from the Phillippines?

 

[Glyfair]

How do they know the accused actually put these items up on P2P themselves?
How can they prove that someone didn't hack their PC and take the files, or a friend using that PC, considering how many folks' WIFI networks aren't secured, PCs that are shared, and how many folk's computers are riddled with malware?

From the documents it is clear that the main argument will point to the fact that the watermarks with their name on it were removed. It shows they were trying to hide their identity. Someone who stole it from them would have little reason to do that.

 

[Charwoman Gene]

Pixels of each page are more than 6 millions. Are PCs robust enough for this? Assuming software exists for the job.

Well, seeing as how most PC can complete a BILLION operations/sec* we shoul be okay.

*Yes, this is a gross over simplification.

 

[Arivendel]

What's the point in sueing a guy from the Phillippines?

Because by suing the ~ don't use profanity thanks - PS~ you make the statement that no matter where you are you WILL be charged, prosecuted and forced to pay up if WotC finds you.

 

[Mustrum_Ridcully]

Pixels of each page are more than 6 millions. Are PCs robust enough for this? Assuming software exists for the job.

Of course they are - they need to display the pixels after all.

Going over even 6 Million Pixels is not all that hard.

A 10 Megapixel image (like from a good camera, about 3,000 x 3,000 Pixels, larger than any common monitor display can present you) needs around
40 Megabyte of memory in RAM if fully decoded. (Assuming it contains the information for the Red, Green and Blue Channel and an Alpha Channel for transparancy, using 255 distinct values - bytes - for each channel).

Think about how much RAM you have in your PC...

 

[Mustrum_Ridcully]

Or someone could just buy a hard copy of the book and scan it.

It wouldn't be perfect, but there would be zero chance of traceability. With some of the higher end vision software these days, you could probably even get a decent percentage of the text recognized to be searchable. If someone cared enough, they could manually correct errors and re-proof the whole book.

Did you know that many printers come with a "watermark", too. Each page is imprinted with it - invisibly to the naked eye, but if you print out something, it could be traced back to the buyer of the printer.

Similar things might be available for scanners, automatically watermarking your scans. (I don't know if that actually is the case, though.) What alone is possible is that, if you create an image with some kind of meta data (like JPEG, the most common file type I suppose), it might just contain the information in your meta data. Of course, removing that is relative easy.

There are probably always ways around it. But scanning an book and checking there are no hidden information in it relating to you is a lot more effort than just uploading some random PDF you just downloaded from a publishing website. It's also more work than removing watermarks from an existing PDFs, since the latter is just a little crunch time for your processor, not requiring you to scan in 100+ pages book.