I’d like to celebrate read-an-ebook week with something positive, but I don’t have anything even half-prepared except some thoughts about DRM. That’s bad enough, but I won’t even be arguing against DRM for ebooks. I certainly won’t argue for it — I’m against the idea on principle, so if I thought it through and concluded that it was actually a smart move for publishers I would have to just keep my mouth shut. There’s a facile comparison, though, between digital books and digital music which suggests that DRM for ebooks is as dumb an idea as it is for music (I did it myself a few days ago, which got me thinking about the whole issue). I think that comparison is wrong, and that we need to look more closely at the properties special to ebooks if we’re going to make a reasoned and convincing argument against crippling them with DRM. The first step is seeing clearly what makes DRM so spectacularly unsuccessful for music, and checking whether those conditions carry over to ebooks. That’s what this post is about.

DRM for digital music cannot succeed.

This idea is slowly getting taken seriously, partly due to overwhelming evidence and partly due to the efforts of activists like Cory Doctorow. Doctorow has a nice analogy, which you can read here, showing why DRM is a basically broken idea. Stripped of its rhetorics (which are the point, by the way; if you’re unconvinced by my exposition I definitely recommend his) the idea is this: DRM provider Alice tries to hide a piece of content from Carol (who isn’t supposed to get access to it) while revealing it to Bob (who is). Only no DRM system can prevent Bob and Carol being the same person (a criminal intending to sell copies of Your Band’s Latest Album buys it legitimately; now the DRM has to let him get it on his Bob days but not let him get it on his Carol days).

There’s more subtlety to this idea than Doctorow admits, though. There are a bunch of factors that make digital music particularly un-DRM-able (to coin a barbarism). Here’s a list of factors ranging from highly technological to highly social (the middle entries are sort of a mixture of both):

  • audio output makes pipelining easy
  • cds and the mp3 format are de-facto standards
  • digital music gets played on general-purpose computers
  • enormous numbers of people listen to music digitally
  • there are social assumptions that swapping music is ok
  • there’s a general perception of record labels as exploitative

Let’s take them in order.

Pipelining

Doctorow’s analogy assumes that Bob and Carol both want the same thing from their file. For music that’s a pretty safe bet, because there’s basically a single bottom line: Bob (the legit listener) needs to be able to get the music out; for reasons we’ll come back to in a moment, he almost certainly needs to be able to get it out digitally (not just as audio output). But that means Carol can take that output and pipeline it straight back in as the input to something else: a cd burner, her mail client, a file-sharing program, or whatever. (Actually even just having audio output would let her do this to some extent, but she would lose some of the information in the original, even with the best-quality gear.)

De-facto standards

Cds and the mp3 format are de-facto standards for digital music. That means attempts to crack DRM can be pretty tightly focussed: get the music off a protected cd, convert it to mp3, release it. And cds are pretty much impossible to DRM effectively, because the same physical technology (a cd drive) gets used both for data and for music cds, when they get played on …

General-purpose computers

A cd can’t really tell when it spins up in your computer’s drive whether it’s being pipelined into an excellent music program or into an mp3 ripper. I said above that Carol would lose quality in re-recording from audio output, but if she can put the cd into her laptop she doesn’t even have to do this. She just has to trick whatever DRM system is on the CD into believing that she’s only accessing the data for the purposes of playing the music; once she has the data, she can do whatever she likes with it. And on a computer that lets you install your own software (your laptop or pc, that is, but not your iPod — at least not without some serious extra hassle) that’s going to be very easy indeed.

Enormous numbers of users

I said “not without some serious hassle”; in the case of music, that doesn’t actually matter so much. So many people listen to digital music that there’s bound to be someone out there willing to put in the time and effort to rip Your Band’s Latest Album, even if you make it difficult. Having cds and playing-on-your-pc as a de-facto standard makes this a lot easier, of course.

Social attitudes

The last two points are pretty fuzzy. It seems to me, based on the people I know and the stuff I read online, that these are prevalent attitudes; there might be quite some self-selection involved here though. That said, I see two attitudes: the assumption that swapping music among friends is basically ethically ok, and the assumption that little if any of the money we pay for music gets back to the artists. These attitudes are pretty clearly going to make folk more comfortable with sharing music without paying for it, whether that’s ‘technically’ illegal or not. It doesn’t even matter if the perception of record labels as exploitative is true: the assumption that this is the case makes people more comfortable with ripping them off, since it’s only fair turnabout.

I don’t think I’ve said anything surprising so far. But let’s see how these observations carry over to ebooks.

No pipelining

This is the most important point where the analogy between music and ebooks breaks down. There is no equivalent to ‘audio output’ for ebooks. You can always, without anything more complicated than a piece of cable and the right plugs, pipeline your music if you can play it. The equivalent low-tech solution for an ebook would be photographing a screen displaying each page in turn.

There’s another post hidden in that observation, about typographical standards and the differences between pdf and on-the-fly adjustable technologies like html, mobipocket and so on. But the basic point is: what the reader needs to see is not necessarily what the reproducer needs to get at.

That’s partly because there are as yet…

No de-facto standards

That’s not really true: there are too many de-facto standards, which is as good as having none at all. There are a plethora of physical reading devices out there, all with different screen dimensions and technical qualities (e-ink, backlit screens, link-based navigation and what-have-you). Content that is right for one won’t necessarily be right for another — both in the trivial sense that a file for the Amazon Kindle won’t be readable on a Sony Reader, and in the less trivial sense that a pdf carefully formatted for the dimensions of the iLiad won’t be pleasant to read on your netbook or laptop.

The lack of standards also means there isn’t a central clearing-house for DRM-cracking efforts the way there is with music: rip the cd to mp3 and you’re done, but get ahold of a pdf and you have to convert it to half-a-dozen other formats before you’ve satisfied everyone. Tools like calibre are going to change that, eventually, but they haven’t gotten there yet.

Special-purpose devices

Part of that lack of standardisation comes from the variety of physical devices being used to read ebooks. The most important of those devices are dedicated ebook-reading systems like the Kindle or the iLiad. Amazon’s Kindle is a closed device: you’re not supposed to be able to run your own programs on it.1 And, contra Doctorow, a DRM scheme based on such a device can work. Again, it won’t work for music, because even plain audio output can be pipelined; if you make the output quality too bad to use for re-recording then the device is no good for listening to music any more. But that’s not the case with an ebook reader: the output is an image on the screen, and there’s no way to turn that into something as versatile as the original file was if the device doesn’t let you get access to that file.

Limited numbers of users

If enough people were using ebooks, even this wouldn’t matter. We would have pirated editions consisting entirely of photographs of authorised editions. It happened to Harry Potter, it could happen to everyone!

Well yeah, except that it won’t. There are too many books and too few people who could be bothered to perform this incredibly tedious work. (Unfortunately for DRM, incredibly popular titles like the Harry Potter series will continue to attract this kind of insane effort, and those are exactly the ones that ought otherwise to give the largest returns for the publishers. But it’s not something for them to be worried about in the general case.)

Different attitudes?

Finally, many attitudes to books are rather different at this point to attitudes to music. This is again pretty fuzzy and probably extremly personally biased. It’s also very much based on attitudes to paper books, which might not carry over to ebooks at all. For instance, paper books are for lending; you lend cds too, but the cd you lend can be copied while the paper book can’t.2 I doubt this attitude will carry over to ebooks, because by now we’re completely used to the idea that data can be copied: I won’t want to ‘lend’ you my ebook because there’s no non-DRM reason why I should be deprived of it while you’re reading it. Again my personal feeling (which is probably not representative of any significant market) is that publishers don’t get the same assumptions of evildoing that record labels do. Here I don’t see any reason why electronic publishing should change matters; on the other hand if major publishers get DRM-crazy and end up associated with the likes of the RIAA, their credibility is going to take a hit (at least, and to belabour point one final time, among the people who are already writing and reading blog posts about DRM).

What does it all mean?

So I hope you’re convinced that ebooks are not like digital music, and that DRM being completely braindead for the one does not automatically mean it will be braindead for the other. In particular, the fact that you order your Kindle ebook through the Kindle, download it using the Kindle’s own network connection, and view it on the Kindle screen without ever having access to the file itself, means that DRM is technologically feasible on that platform. (Bob gets to read; Carol doesn’t get to copy.)

Yet there is hope! There are so many hardware options, and so many competing standards, that restricting to Kindle-only (or Sony-only, or…) for the sake of DRM will be a bad business move for publishers unless Amazon manages to strangle all the competing alternatives before they get serious market share. So long as you can get the original DRM-locked files onto a general-purpose computer, the DRM is eventually going to collapse: Bob needs access, but Carol can dress up as Bob. So we need to lobby for the ability to share ebooks across multiple devices, to make our own backups, whatever else gives us an excuse to get the file out of that device and onto our own computers. If ebooks are sold under the assumption that this is possible (and if enough people are reading them to make cracking the DRM a fun hobby for someone), these DRM schemes will eventually crumble.

What still has me seriously worried, I have to admit, is that I don’t yet see a sensible alternative to DRM-based restrictions that still gives me high-quality ebooks to read and that will be attractive to publishers. A music junkie who will fill 150G with mp3s is going to carry on beyond 150G. I’m not sure that’s the case with books: one can only read so much in a lifetime, and 150G might be pushing that limit. A swapping session with a new friend might give me a couple of days worth of new music, including a couple of albums that send me looking for more stuff by the same people. The same session with someone with a nice library might mean that I never have to buy anything by Gene Wolfe: not many people have complete album collections for any but their favourite bands; I bet though that everyone who reads beyond the first book of a series will accumulate all the others if they’re easily available. It could keep me reading for the next two or three months. That has to be a scary thought for publishers.

The most obvious alternative to DRM is simply to keep some titles non-digital, or to provide them but in deliberately low-quality versions to give an incentive to buy ‘the real thing’. I don’t like that alternative any more than I like DRM: I want my ebooks to be of high typographic quality. Next ebook post (which almost certainly will be after read-an-ebook week is over, perhaps long over) will be about this issue, and (again rather depressingly) about how often this is not the case: demonstrably so at present, and I think likely also for quite a while to come.

Notes:

  1. The iLiad, on the other hand, is largely open-source. I own one and it’s a fun thing to play with, as well as good for reading books on. Go the iLiad! Apart from the name, which is quite frankly embarrassing. [↪]
  2. Yes, it can, but not easily and the result is not nearly as pleasant an object as the original. It still happens among students for textbooks, though, that’s for sure. [↪]