The aim of an End User License Agreement is, nominally, to present the user with the details of legal rights and responsibilities of both the user and the software manufacturer, set out as terms and conditions of installing and using the software. It is usually presented to the user of a computer during the installation of a piece of software in the form of a text box containing the text of the agreement. A method for the action of consenting is given, usually a check box that requires the user to click it to ``agree to'' or ``accept'' the license agreement, and then some sort of button that allows the user to proceed to the next stage of the installation process. This sort of consent would be best described by the effective consent definition of consent given by Faden and Beauchamp (1986) 1.2.2, since the consent-requester discloses and ostensibly presupposes competence2.1. This sort of consent appeals to a justificatory sense of rational autonomy, in that it attempts to respect, to a certain degree, the autonomy of the computer user.
There are some fundamental problems with this style of agreement. With the user only able to see the terms and conditions of the installation of the software once they have begun the installation procedure, the user is often unable to assess these before purchasing. This is especially true of software purchased from a shop, which is why EULAs have often been termed ``shrink-wrap licenses''. The shrink-wrap license is problematic because shops that sell software will not usually accept returns of the software after it has been opened due to piracy concerns, regardless of the reason for return. One such example is EB Games, which states it will not accept a return unless it has ``all original packaging intact'' [EB Games, 2009]. The possible exception here is under trade practices legislation, which, in Australian law, provides for return of faulty goods, goods which have been improperly described, goods which were different from the sample shown, or goods that do not do what they are supposed to do [Commonwealth of Australia, 1974]. Unfortunately, though, it is difficult to know which area covers rejecting an End User License Agreement. Even then, many people believe that simply opening the shrink-wrap covering the purchased software box means they've implicitly accepted the EULA within, and might find it intimidating to go through the process of returning software if resisted by the vendors. Thus, as Erik J. Heels laments, ``they already have my money, I'm clicking accept'' [Heels, 2004]. There have been a few high-profile cases of users fighting against this practice, such as with the OEM-bundled versions of Microsoft's Windows operating system [Bennett, 1999], but still today the norm is that software returned in this way is not generally accepted back by manufacturers or vendors after opening. This also applies to software purchased online, but in many cases there are trial versions of the software or accompanying Web sites that detail the EULA before the software is purchased and downloaded. A study by Carl Belgrove (2008) showed that 14 of 25 surveyed software products did not mention on their packaging that installation of the software was contingent on accepting a license agreement. This practice thereby undermines the consenting procedure, since it contaminates any kind of motivation for really freely consenting - the user has paid their money, so they have financial incentive not to fail to consent.
Another problem lies in the length, language, and presentation of the EULA. As I show in the examples, and in Figure 2.1, the agreements are often thousands of words long, and written in a legal style of language. Figure 2.1 shows this particularly well, since the scroll bar on the text box is quite small, indicating a large amount of text in the text box. The use of legal language is understandable, since the agreement is, essentially, a legal document that discloses terms and conditions in a legally binding manner. However, this makes it very difficult to read and understand by the users who wish to install the software. The length of the agreements also makes it difficult to read, since computer screens can only fit so much text on them before requiring the user to scroll down to read the next page. This, combined with the problem that the user interface of the EULA acceptance dialogue box usually squeezes the agreement text into a tiny part of the screen (as seen in Figure 2.1), means that the overall experience is not aimed at helping the user to read and understand the agreement; rather, the EULA is a way for a software company to disclose their legally obligatory terms and conditions with very little thought for the user who is trying to decide whether they should accept them. This has become standard practice throughout the software industry, essentially institutionalising poor informed consent procedures that alienate and disadvantage the user while maintaining enough of a claim of informed consent to satisfy legal institutions. This is supported by Belgrove's study that showed that companies mostly designed EULA screens to suit their purposes, rather than designing them with the user in mind, which puts users at a significant disadvantage when it comes to reading and understanding their rights and responsibilities [Belgrove, 2008]. An example is shown in Figure 2.2, an EULA for Cakewalk Music Creator 3 which Belgrove describes as ``...set out in a way that protects the interests of the provider, by highlighting only certain elements important to them. This failure to present a full summary of the terms suggests it was not designed with the consumer interest in mind'' (p. 9). There are several studies that have shown that computer users do not actually read the contents of the EULAs, because of the above reasons [Good et al., 2006,Magid, 2005]. In one of these, a software designer tested this (somewhat facetiously, but the point stands) by including a clause in the EULA that offered a reward for writing into a particular email address. After over three thousand downloads and four months, one person finally wrote in to claim the reward [Magid, 2005]. Another study showed that users, although understanding they were entering into a contract, did not read the contents of the EULA, and when told the contents of the EULA after installation regretted installing the software [Good et al., 2006]. Further studies show that even if users did read the EULA and discovered some undesirable content within, they are likely to consider how the immediate benefits (instant gratification, completion of a task) outweigh the undesirable side-effects [Acquisti & Grossklags, 2005], especially when presented with ambiguous or uncertain language.
However, even the satisfaction of legal standards is in question with EULAs, since there has been no court challenge of the validity of a license agreement as a whole, only challenges to various parts of license agreements, which have been found in favour of both sides, with some of the terms being declared unconscionable or in other ways legally unacceptable contracts, such as Klocek v. Gateway, Inc., and SoftMan Products Company v. Adobe Systems Inc. (also [de los Reyes, 2002]), and some decisions stating that the licenses are enforceable and valid, such as ProCD, Inc. v. Zeidenberg [de los Reyes, 2002]. Although I wish to avoid delving into the depths of the related law in these cases, it seems that the courts (at least the US courts, where most of these challenges have been made) generally fall back to other areas of law in order to deal with license agreements, mostly contract law, which deals with expectations that users will act autonomously, and deal with the agreement in a similar way to that of any other contract (despite the differences between these types of contract, in frequency and manner of consent, at the very least). Clicking the ``I agree'' button in the EULA acceptance dialogue box is considered enough for the user to be held to the terms of the agreement, under US law, as in ProCD, Inc. v. Zeidenberg, because the license dialogue box usually does not allow the user to proceed with the installation without ``indicating acceptance''. Although contract law may be appropriate for the legal basis for challenging the contract of the agreement, there has been little legal testing of the nature of the agreements themselves, despite the differences between IT and non-IT contexts. Belgrove also identifies other key issues with the legality of EULAs, such as frequent reference to other countries' legislation, ambiguous references to things such as ``statutory rights'', and excessive exclusion of manufacturer liability, which then often refers to the above legislation or rights, further miring the user in uncertainty as to what their actual rights are [Belgrove, 2008].
The examples below illustrate these problems in detail. Firstly, they show the poor attempt at implementing any sort of informed consent procedures, but also how the Faden and Beauchamp effective consent model (see section 1.2.2) is used by default, but without any competence testing or assessment of understanding. The case studies also show some of the resulting issues that have arisen from these poor practices, and how these are ultimately poor standards based on an informed consent model that is inappropriate for information technology.
