Research Motivation

As we being above-mentioned, the two concerns--how to collect works and how to circumvent potential copyright infringement allegations--are very important to the Internet library constructor. The foregoing examples demonstrate several strategies adopted by the website constructors in respect of these two essential concerns. In terms of the first concern, there are two choices available to the website operators:

one is employing software and another is collecting works manually. As to the copyright issue, specifically examining copyright to make sure how he can use the works is one option; another option is relying upon the copyright exemptions. In fact, an Internet library is a website from the users’ viewpoints. That is to say, an Internet library constructor may adopt the strategies similar to the website operators. Therefore, if we focus on “employing software robots to collect works” as well as “examining the copyright” and use these two as the vertical and horizontal axes, the websites can

1 http://en.wikipedia.org/wiki/Project_Gutenberg Non-exam copyright

Non-Robot Robot

Exam

I II

III IV

ex: IA ex: Scribd

ex: ACM

be placed in the one of the four quadrants in the following diagram:

Figure 1: Libraries in the four quadrants

In the first quadrant, a library (Model I Library) relies on the traditional library exemptions to avoid potential copyright infringement allegations. The second (Model II Library) and the third (Model III Library) both depend on the licensing of authors, but the Safe-Harbor exemption is more important to Model II Library the because the libraries of this model, at most times, do not explicitly monitor the correctness of the licenses it obtained, rather disseminating works in good faith. Moreover, libraries merely focus on the materials of the public domain should be placed in the third quadrant as well.

In light of the various strategies, the risks of copyright infringement allegations are different as well: not surprisingly, libraries belonging to the first and second models have the highest risks; the reason will be rendered in the following sections. On the contrary, the risk of a model III Library is relatively low. However, in the real world, the lower risk is not free at all and, in fact, is relatively expensive: As to a library of the third model, the time and money spent in completing the negotiations between the publishers and authors are quite significant. On the other side, the cost in respect of confirming copyright authorization scopes of the other two models are relatively low:

libraries of the first model do not pay any attention on it and, libraries belonging to the second model almost pay nothing neither because a Model II library only removes works whenever it receiving notices.

Apart from the concern about the copyright infringement, another important concern is about the way to create collections in the library. As we have seen, the libraries of the first one model clearly face a higher legal risk than libraries of other three models. In general, the reason of taking such high risk is that, subject to the same budget, the total amount of collections in a model one library is higher than the other three models and, at the most times, the amount of collections is one of the most critical issues to a library which may actually affect the users’ favors. The reason why a library of the first one model can acquire more works than the others is that it

employs software robots to collect works on the Internet. In respect of the huge number of collections in the libraries of the first model, fair use, or other general copyright exemptions, is the only effective way which libraries of this model could account on because the total number of works is too massive to explicitly identify the scope of copyright authorization.

On the other hand, libraries of the other two models collect works without any software robot. Since a library of the model II depends on favors of the contributors or authors, the constructors of a library of this model can not passively decide the total amount of its collections; as a result, in general, the total amount of collections in a Model II library is less than it in a Model I library.

As to the other the third model, the amount of collections of a library of this model is relatively limited, because it collects works by hand and, in general, the human’s work speed is less than an unstopped software robots. For example, in spite the amounts of collections in some present libraries, such as ACM Digital Library (ACM Digital Library, 2009), are relatively large; however, comparing to the total number of works on the Internet, the collections of a library belonging to this model are still relatively limited because such libraries have to be subject to their budgets. A summary of these three models are also shown as follows:

Table I. A summary of the four models in risks, costs and amounts of collections

Copyright Policy Risk Cost Number

I General copyright exemptions: fair use etc. High Low Almost unlimited II Licensing from authors and the Safe

Harbor exemption

Medium Low Limited

III Licensing from authors Low High Relatively

limited The Model I and Model II libraries both depend on copyright exemptions, however, the traditional library exemptions could not directly and clearly apply under this circumstance since the conditions are not satisfied (Bolin, 2006). Furthermore, great diversities appearing in the copyright limitation and exception rules in different

national laws increase the risks. For example, the scope of “fair dealing” in the UK is much narrower than “fair use” in the US, as the former has no general exception of the later (Cornish and etc., 2003a). Moreover, the “private use” exceptions in the civil law countries is much common than it in the common law countries, as the civil law countries respect the intelligence in the work rather than the exploitation benefits in it.

On these two grounds, the Internet libraries can not firmly rely upon the limitations and exceptions to lawfully access to, reproduce, even redistribute as the exceptions of individual national legislations are diverse and, under some circumstances, unpredictable. Even though, ignoring the diversities and uncertainties of copyright limitations and exceptions, the copyright exceptions could be applied, the ‘fair use’ or other similar exceptions inevitably undermines the quality of contents because the future uses of the contents are bounded because users of the library can not be sure what the exact authorization scope of the work is.

On the other hand, the simplest solution to reduce such high legal risks is to explicitly examine the copyright of each work, such as getting license from the authors or only collecting public domain works. However, a specific analysis of the copyright of a work is very difficult and needs a lot of human and financial resources.

As a result, the number of the libraries belonging to this model is quite limited.