7 June 2012

A few random points on DRM

What follows is my response to the IDPF's EPUB content protection proposal.

DRM works, in some contexts such as software.

One of the advantages of software and apps is that you can apply different forms of copy control for each product or even for each release, catered to the needs of each product.

Big budget games have extremely front-loaded sales (most of it in the first few weeks) and so heavy-handed DRM actually makes economic sense. It doesn’t have to last forever, just over the initial sales bump.

An app can try lesser or greater forms of copy controls in each release and do a detailed analysis of which one is the most cost effective and which is cheaper.

A long-standing title can change the copy controls in the next release if the early one is broken.

Software defines its own behaviour and, as a result, they have much more flexibility in how they apply copy controls.

The problems with applying DRM to ebooks is that they don’t work in the same way as software.

Point 1: Ebooks aren’t software and, as such, can’t use DRM effectively

Ebooks require standardisation of some sort, whether it’s within a single supply channel or across the industry. A publisher cannot experiment with copy controls from title to title because the structure of the industry and the nature of the ebook (it’s a file not an app) means they have to pick a standard DRM method.

The problem is that standardised DRM is ineffective. It has none of the pragmatic advantages of software copy controls, but all of the problems. It can’t adapt to the needs of the title (books with front-loaded sales versus perennials need different kinds of DRM if you are going to go that route).

Which brings us to…

Point 2: Standardisation directly counteracts the advantages of DRM

Standardised DRM loses the adaptability that is the only reason why DRM survives in the software industry. To get the same theoretical advantage from DRM as software does, requires a proprietary integrated system. If you are going to try to use DRM to prevent copying, pirating, and sharing, a proprietary system like Apple’s or Amazon’s where they control every stage of distribution is the only one adaptable and flexible enough to respond to changes and events.

(That they don’t just furthers the point that, to these companies, DRM is just the cost of doing business, not a strategic imperative. They implement DRM because publishers ask for it. If they cared about its effectiveness, they’d be changing and adapting the DRM schemes every week, at a great cost, and to little benefit. No software company will ever have the economic incentive to do content DRM properly and only software companies have the resources and talent to do so.)

Point 3: DRM compromises interoperability and modularity

The very nature of DRM and copy controls make open and modular systems more expensive and complicated to implement. Creating an open standard for DRM doesn’t solve this problem as it means that every other stage in the retail chain has to be changed to reflect the new DRM standard.

Which brings me to the corollary to point 3: Standardising DRM will have wide-ranging consequences on other standards and will both delay them and make them more complicated.

Point 4: Movies and video are a lousy role model

If it’s on video, it’s on the web. Video DRM technology is quite possibly the least effective forms of DRM there is.

This is blatantly obvious to even the most casual observer. It doesn’t matter if the video has been broadcast on TV, streamed on the net, or released on DVD or Bluray. The video is on the web, often within hours. Streams, DVDs, and Blurays are trivially ripped even when encrypted.

DVDs are, at this point, the perfect example of ‘lightweight’ DRM, and they are also the poster child of pleasantly useless copy controls. They are a clear demonstration that the idea that lightweight DRM would help publishers prevent sharing and pirating is ludicrous.

Moreover, anti-circumvention laws that are designed to prevent these rips haven’t had any effect on them whatsoever and apps for this purpose remain easy to find. (More on the anti-circumvention laws below.)

Point 5: The economic impact of piracy cannot be measured, but the impact of DRM can be measured

In fact, it’s trivial. If adding DRM to ebooks increases sales, it’s working. If removing DRM from a title decreases sales, it’s working. You take a large group of titles with known aggregate performance, remove the DRM of half and keep it on half and see what effect it has on the sales in each group.

The only thing that matters in this discussion is the ROI of DRM. Piracy is a different subject. It doesn’t matter what you think of piracy or sharing, we live in a capitalist society and these are publicly owned corporations, a decision such as the one to use DRM should be backed by cold hard numbers. If it makes money, keep it. Otherwise you’re a fool if you don’t lose it.

If publishers don’t have the numbers to prove the economic value of DRM, then they have no right to ask the IDPF to standardise it because they can’t give you any facts or context on which to base your design.

Point 6: Publishers could be more flexible about DRM if they wanted

The irreversibility is a myth. A publisher could easily go back to DRM if they wanted. Sure, they can’t reDRM sold titles, but any published title has to be treated as if it is DRM-free, because of how trivial it is to remove it. Pretending that there is an issue with experimenting more with DRM-free is either disingenuous or just plain dumb (pick one, or both, if you aren’t feeling particularly generous).

The only valid reason why you can’t go back to DRM is because of the inevitable customer outrage, which is a sign that DRM was a bad idea in the first place.

Point 7: DRM is inherently consumer-hostile and limits speech and expression

As much as I appreciate the economic role of DRM and copy controls in the software industry, Content DRM introduces too many issues for the reader.

It makes fair use more difficult.

It often limits quotes and disallows otherwise legal and reasonable use cases.

It is also a blatant act of consumer-hostility: the publisher is plainly stating that consumer can’t be trusted.

While in actual fact, because of the ineffective nature of standardised DRM, it only affects honest people. Crooks either strip the DRM or just download a DRM-free file from a pirate site.

Ebook DRM is a clear case of corporations being nasty to their customers.

Point 8: Standardising an ineffective solution is dishonest

No technical body with integrity should even contemplate standardising an ineffective solution just to leverage a legal loophole. Lightweight DRM is an oxymoron. It either prevents piracy and sharing or it doesn’t.

DRM should stand or fall on its own technical merits. Piracy is already illegal. You don’t need additional legal tools, tools which are too often used to silence critics and cover up corporate misdeeds. (These are very bad laws.)

Anti-circumvention laws are useless at preventing the production and distribution of DRM-stripping software (finding apps like, for example, Handbrake or VLC is very easy).

They do not prevent pirate sites from hosting ripped files.

What anti-circumvention laws do is help lawyers cover up corporate malfeasance.

One friend of mine refuses to set foot in the USA because he knows that he has broken these anti-circumvention laws as a part of his job several times and he doesn’t want to risk getting into legal trouble. He is an idealist working in software security to make the online world a safer place, and he’s much more likely to be a target of these laws than any ebook pirate. It’s a law that lets corporations target white hats who discover uncomfortable facts.

These are very bad laws, wherever they have been in effect, and proposing an ineffective (lightweight is just another word for completely useless) copy control scheme just to leverage these laws is both dishonest and in very bad taste. Publishers should be working to repeal these laws, not build their business models on them.

Point 9: You have bigger fish to fry

The FXL spec needs clarifications . (I have to say that the conclusion to that discussion disappointed me.)

Then there are problems with javascript in ebooks. Before we can even think about wide, cross-platform, uptake of javascript in ebooks there needs to be work done on several issues with javascript security in ebooks. Apple clearly understands these issues (the limitations they have set on normal EPUBs and on iBooks Author books are pretty much exactly what I would have recommended). There is a lot of work that needs to be done here. My earlier post on issues with javascript in ebooks doesn’t even begin to scratch the surface of this issue.

Finish the dictionary and index work.

And there’s more, I’m sure, this is just off the top of my head. There’s lots to do. No need to add this to the list.

Point 10: Cultures where piracy is endemic won’t be fixed by DRM

If you’re screwed, you’re screwed. If proprietary DRM isn’t preventing piracy, standardised DRM won’t do a thing. If existing DRM doesn’t give them enough legal tools to combat piracy, a standardised solution won’t help them one jot.

Only one file needs to be stripped of DRM and uploaded for the whole copy control endeavour to be pointless. It’s an ineffective waste of time, energy, and money. If these companies live in a culture that no longer values cultural production, no amount of IDPF hand-waving will save them. They are screwed.

Point 11: Corporate encryption

Corporations don’t need lightweight DRM. They need proper encryption. You can’t do one solution that is both lightweight and industrial-strength. You can’t do both. Either do a proper, industrial-strength, encryption scheme that would satisfy even the most paranoid corporate IT security goon. Or you don’t. Creating ineffective fobs for corporate lawyers to hang their lawsuits on is beneath the dignity of any honest technical standards body.

Besides, EPUB hasn’t been integrated into the corporate document workflow and it’ll be hard to dislodge PDF from it. There is also no point to it. PDF is doing that job very well. Corporations gain precious little from EPUB and we gain nothing from compromising EPUB to suit their purposes. Especially since they won’t use EPUB anyway, for workflow reasons, not encryption reasons.

Final point: You’re going against history

Now, normally I don’t put any trust in the toxic mythology of inevitable progress but content DRM is quite simply going to fade away.

No matter what your opinion on piracy is, or what attitude you take towards fair use, content DRM is utterly useless.

Useless things die out.

Just let them.


What follows is Bill McCoy’s comments on my response to IDPF’s EPUB Lightweight Content Protection proposal:

Thanks for the response; your comments are well taken but I’d like to respond to some points, in the interest of a constructive conversation (what follows is only my personal opinion).

First, I completely agree that “eBooks require standardisation” and that eBooks are more content than app . The lines are blurring (“The Elements” is an eBook packaged and sold as an app, with app-like interactivity) but generally speaking eBooks can be categorized more with music and movies than with software or games. I also agree with you that strong DRM is only possible with apps, and then only for relatively short release windows (i.e. the hackers will eventually strip any DRM). I further agree that standardization of DRM directly counteracts the ability of DRM to withstand determined hackers (even for brief periods). So, I feel that you and I share a lot of common perspective.

But to me these points do not at all imply that standardized DRM is useless. The locks on most of our homes’ doors are standardized and trivial for a determined burglar to quickly defeat. But most break-ins are the teens down the road and others who aren’t determined burglars, and the point of locks isn’t really to keep out determined burglars, so much as a reminder of a social contract. You say that only software companies do DRM “properly” but that’s like saying only banks do locks “properly”… it may be true, but it’s a rather meaningless statement.

The operative question to me is whether book publishers should aspire to the position of movie studios, whose DVDs have a standardized (and thus easily breakfable) DRM or music labels, whose CDs can be copied by legitimate software. Well, if I were a book publisher I certainly would rather be in the studios’ shoes than the labels’ shoes! Both formats are widely pirated (as you point out if it’s on video, it’s on pirate sites) but based on a sample size of one 15-year-old son I’d argue that casual sharing of tracks ripped from CDs is much more prevalent than casual sharing of digitized DVDs. One datapoint is that Walmart announced this week a program to allow consumers to get digital versions of DVDs for a couple bucks. Regardless of whether you think this is a valid reason to extract incremental revenue from consumers, it’s clear that no one would pay a dime for a similar service for CDs, because they can and do just do it on their own. That’s prima facie evidence that DVD DRM, weak though it is, is not useless.

I agree with you that the utility of DRM in reducing piracy/over-sharing is debatable. But it’s simply a fact that many (presently, most) authors, publishers, and distribution channels require DRM on eBooks. And, a number of them have asked IDPF for help in standardizing it. It is also a fact that the proliferation of silos of non-interoperable DRM reduces interoperability for readers, making it harder for them to read when and where they want. This is counter to IDPF’s mission to foster EPUB as theglobal interoperable open standard.

I also agree with you that the best that we can reasonably do as a broad standard, based on both the nature of eBooks and the costs that publishers are willing to bear, is something pretty weak by comparison to the DRMs used with software and games, something more like DVD DRM. Determined pirates won’t be deterred, nor will those determined to patronize pirates. But with no DRM at all (a la CDs) the chances of a book group, or a college dorm, all sharing one copy are arguably much higher than with a DVD-level DRM. Because, today these types of groups do share tracks ripped from CDs and generally don’t share digitized movies. And as a reader and eBook-buyer I’d much rather have a single lightweight DVD-level DRM than a plethora of more draconian DRMs to deal with. I’d rather still have no DRM at all, but that’s not a scenario I see as likely to happen. So, a lightweight DRM seems to me (again, personal opinion) a reasonable balance of publisher and reader interests and cost. Like the “improper” lock on my front door.

This applies as well to your statement that “Corporations don’t need lightweight DRM. They need proper encryption”. 99.9% of encrypted documents published by corporations are PDFs with basic password protection, as per for example what’s created via the system print dialog’s Security Options on Mac OS/X. Are these truly secure? Absolutely not! In fact without an “open” password, choosing to disable printing & copying is really only obfuscation: the algorithm to decrypt is publicly available, and there’s no legal requirement for PDF readers to honor these permissions. But the point of the prevalent lightweight PDF encryption isn’t to protect truly confidential data or ward off determined thieves, it’s just to keep basically honest people honest. EPUB is starting to replace some uses of PDF in corporation, as EPUB improves N-screen support, accessibility, and integration with Web Standards. It doesn’t seem unreasonable to enable the similar (weak) level of encryption to remove a barrier to switching.

As far as whether it’s a priority for IDPF to work on any of this, that will depend on the feedback received from members and others. As you point out, the IDPF has a fair bit of other work on its plate. And, assessing likely adoption of any solution we developed will be a critical factor, because if we just add an N+1 DRM scheme that is not broadly adopted anywhere, then that would end up hurting not helping our interoperability goal. I’m hopeful that we have a window of opportunity to head off increased fragmentation chaos, at least in some regions, but that’s TBD. I do think though that it’s “now or never” given the transition from EPUB 2 to EPUB 3, the imminent growth of eBooks into a significant business outside the English-speaking markets, and vendors like Kobo starting to defect from the de facto interoperable solution (Adobe ACS4). The spirit of the IDPF draft document is in the nature of “if you want us to do something, here’s what we think we could do”. Whether we do it or not is up to our members. But even a year from now, I think it will be too late. Not (regrettably) because eBook DRM is going away, but because the vendors in every region will have rolled their own DRMs.

Now, I don’t think anybody would be surprised to say that I disagree with quite a few of his points. In ebooks and movies we are living with a software-encoded honour system where the limitations on the consumer’s actions aren’t in any real way enforced; they’re just guided through software-encoded protocols. The encryption, because of how trivial it is to break, is incidental to the process. The only real enforcement in the anti-piracy battle is performed by the lawyers (the foot soldiers being marketing and sales staff). How many of those sued for piracy of DVD-ripped movies are sued under anti-circumvention laws?

None. That’s how many. Encryption is a non-factor in the battle against piracy. It doesn’t prevent it and laws regarding encryption don’t have a role to play in enforcement.

The fact that I can link to VLC without any repercussions and the fact that it is one of the most popular video playing apps today demonstrates how ineffectual DVD DRM is and how useless the anti-circumvention laws are in preventing the creation of these tools, internationally. Those laws are, on the other hand, used against researchers who point out just how ineffective Adobe’s PDF encryption schemes are. De facto, the only purpose of anti-circumvention laws is to limit free speech and fair use.

And even if I weren’t against content DRM I’d still be against this plan. For it to be a priority as a part of the IDPF’s work, they have to have faith that it would at the very least completely replace Adobe’s scheme in the market.

I really doubt they will succeed. If this goes forward, we’ll just end up with one more DRM scheme and suffer more fragmentation in the market. Adobe won’t let go of its cash cow without a fight. Neither Apple nor Amazon are going to switch away from their schemes. The best case scenario is that IDPF’s scheme forces Adobe to change the pricing and business model of it’s DRM system. For that reason alone I suspect that a large majority IDPF’s members is in favour of going forward with standardising this DRM scheme.

My guess is that it’ll be standardised very quickly, used as leverage to improve members’ deals with Adobe, then be summarily dropped.

And it’ll stand on IDPF’s website as a sad, unused, reminder of the industry’s misplaced priorities.

Everything on this site is written by . In case you hadn't guessed already.