A Deriving Quantitative Similarities to Computer Programs and Determining Copyright Infringement in South Korea

Abstract

A substantial similarity is one of the factors for recognizing copyright infringement. Determining whether there is a substantial similarity is a hard work. For computer program works, judging the substantial similarity is even more difficult, because professional ability is required to measure. In Korea, this problem is solved through the expert testimony system in the trial process. In trial practice, quantitative similarity is mainly used as a method for determining substantial similarity. However, there are several limitations in two methods taken here, namely the use of the expert testimony system and the use of substantial similarity. Appropriate responses are required to overcome this limitation.

1. Introduction

When a trial is conducted because of copyright infringement of a computer program, a problem has arisen. The problem is that judges cannot easily understand the expression of computer programs, which are the subjects of a lawsuit. In order to judge whether copyright is infringed, the determination of facts must be preceded. But the judgment for this is bound to be limited because of the judge’s ability of perception. This phenomenon is also a characteristic of modern lawsuits that are increasingly complex and specialized. The recent expansion of modern lawsuits has caused a change in the role of judges in the trial process. There are some limitations of the lack of judges’ knowledge in the professional field. So, in particular, the systems of utilizing experts for trials are gradually developing to overcome the limitations. In addition, this change is actively shown in intellectual property disputes, especially in the lawsuit of copyright infringement with the aim of computer programs.

In the Republic of Korea, the appraisal system [1] as a judge’s order appraisal is widely used in litigation procedures. The system reflects the ex officio principle. Therefore, in using this procedure, several discussions on how to utilize experts and how judges accept the results are taking place in practice. There is a method used in the case of the copyright disputes of computer programs in Korea. It is deriving a degree of similarity with numerical values is broadly used to determine whether the copyright is infringed. Also, judgments based on the derived similarity is used either.

In general case, experts only analyze and identify objective facts without the intervention of subjective values. However, computer program copyright infringement lawsuits are characterized by experts requiring a value judgment to determine the facts. Separate from the judgment of ‘same’ or ‘different,’ the concept of ‘similar’ is bound to be reflected in the above numerical values indicating the degree of similarity. It inevitably involves subjective judgment to some extent by the subject who judges it. At this moment, this subjective judgment is fully based on the experts’ discretion. But the important thing is that it should be handled under the theoretical limitations perceived by copyright law.

Such software forensics plays a critical role in copyright litigation in Korea. Therefore, many discussions have taken place regarding copyright law or litigation procedures. Hereinafter, we will examine the overview of the appraisal process, which is a system where software forensics can be utilized in copyright infringement lawsuits in Korea. So, we’ll see how software forensics are being used.

2. Copyright infringement in computer programs and judging practical similarity

2.1 Types of copyright infringement in computer programs

Disputes over copyright infringement on computer programs occur because of various causes, such as transaction relationships of computer programs, turnover of computer program producers, contractual relationships, etc. After analyzing various cases, copyright infringement in which practical similarity becomes an issue can be classified into the following four types.

Type 1. One unfairly acquired a computer program and used it to create the same kind of new program.

This case is the most basic types of disputes over copyright infringement on computer programs. In this case, the infringer defends himself/herself by mainly arguing that the computer program developed by the infringer is completely different from the one owned by a victim of the infringement. Another case is that if the victim of the infringement developed the program by referring to the technology of external third-party company. The infringer argues that the program was developed by referring not to the victim’s but to the same third-party company’s technology [2].

Type 2. One moved to another company and created the same kind of new program using source code from the previous company.

This case is also a very distinctive type of dispute. However, there are various reasons why the infringer possesses the original copyright company’s source code. One of the main types of leakages is duplicating the source code during the development process to work outside the office, especially at home. Another one is the case of directly duplicating and taking the source code when leaving the company [3].

Type 3. While one was legitimately using the computer program, he reversely engineered it and created the same kind of new program.

Though it is not relatively common, this type mainly occurs in web-based software that can easily reverse-engineer the original program and requires a quick business response. It happens wherein a user uses the program by paying a certain cost first. Later, the user creates and uses a new computer program with the purpose of avoiding usage fee payment. Since, it would be easier to create another program by modifying the original one [4].

Type 4. One created the same kind of new program using computer programs from other companies while he acquired in the course of his work.

This type corresponds to a large-scale and complex computer program compared to Type 3. It starts with a relationship that helps copyright owners based on a professional level of understanding of the original program. But after that, the infringer takes away the market of copyright owners by engaging in unfair copying [5].

2.2 The structure of judging the copyright infringement

The Copyright Act does not specifically stipulate principles capable of judging acts of copyright infringement by law. But considering the basic principles of copyright law, certain standards by theories and precedents are formed and used. Even if they are the criteria for determining whether copyright infringement occurs, opinions on it are not fully consistent. In general, three criteria are applied to the precedents and theories in Korea. First, the plaintiff has valid copyright; second, the defendant created the work of infringement after seeing the plaintiff’s work with copyright; and third, an identity or substantial similarity exists between the defendant’s work of infringement and the plaintiff’s work with copyright. Copyright infringement can be recognized when the above conditions are all met. Of the three criteria above, the third issue of substantial similarity is the most important. Compared to issues wherein a relatively clear conclusion can be made according to specific facts, the question of similarity is inevitably a value judgment made by a judge. In other words, in a state that they are not completely the same or different, the question of how similar two works can be viewed as similar to each other. The criterion for making a judgment is that the degree of similarity that confirms copyright infringement must be substantial. Deciding upon that degree is a very abstract and difficult matter. It sets a boundary somewhere on a straight line connecting the two ends of similarity, completely different and completely identical.

It is difficult to judge whether there is substantial similarity because it cannot be easily judged because of the abstraction and uncertainty of the degree of similarity. In practice, therefore, the judgment on whether copyright is infringed often depends on whether substantial similarity can be recognized.

2.3 Limitations of substantial similarity judgment and calculation of similarity (%)

It is indeed practically impossible to set clear and objective criteria for determining how similar works must be to be said they are substantially similar. Because of this point, it is believed that the process of determining whether a copyright is infringed inevitably connotes ambiguity.

However, the purpose of the trial in a copyright infringement lawsuit is not to reveal to what extent the two subjects of comparison must be similar to be regarded as substantially similar relationships. But it is to determine whether they are substantially similar when there are specific similarities between them. Because it is virtually impossible to find the point, judges do not find a point of criterion to judge that they are substantially similar in a specific case. The issue can only be solved to the extent that it determines whether a certain ‘degree of similarity’ belongs to the right or left of the point of criterion that would be somewhere [6].

In the process of determining whether practical similarity can be recognized in the copyright infringement dispute over computer programs, in the trial practice in Korea, the degree of similarity is mainly used. The degree of similarity between the two programs is expressed as figures (%). Although this method has some limitations, it is used by supplementing them in practice because of its high efficiency in problem-solving.

3. A judicial procedure for determining substantial similarity: expert testimony

3.1 The concept of appraisal

An appraisal is an investigation of evidence to supplement the judge’s judgment ability. It requires a person with professional knowledge and experience to report to the court. The court gets the regulations and empirical rules, or the fact judgment derived by applying them to specific facts from experts. The Korean litigation law system stipulates expert testimony in Article 335 of the Civil Procedure Act. The expert testimony stipulated in the legal procedure laws means to perform an appraisal by a judge’s appraisal order.

3.2 Procedures of Expert Testimony under the Civil Procedure Act

3.2.1 Overview

The expert testimony of the specialized field, a type of evidence supplied by expert witnesses, the contents are more professional and complicated than judges look. In this case, it is subject to practical limitations in applying the principle of free evaluation of evidence [7]. In some cases, it has led to problems hindering the procedural processes, such as high expert testimony fees, poor and insincere expert testimony, delays in reporting the results of expert testimony, moral hazard, as well as preventing quickness and economic efficiency. Therefore, there has been an ongoing debate on the need to improve the procedures at the level of judicial reform.

3.2.2 Duty of Expert Witness

This regulation considers serious problems such as poor expert testimony and outsourcing — blanket subcontracts — from the expert testimony of construction appraisal in the past. It can be seen as responding more strictly to poor expert testimony. Also, it would be seen as trying to increase the reliability and level of expertise by explicitly stipulating the regulation, which had not been explicitly recognized in the law [7]. It also embodies the application of the multiple appraisal system. In general, it is common for one expert to be designated as an expert witness. But Article 339 of the Civil Procedure Act regulates that multiple expert witnesses shall state their opinions together when they receive the request for expert testimony. Considering that, the multiple appraisal system is recognized under the existing regulation. In addition, when multiple appraisers conduct an appraisal, all of them shall be designated by the court when conducting an appraisal jointly to prevent the entrustment of expert testimony.

3.2.3 Methods of Stating Expert Testimony and Examining Expert Witness

The revision originated from the problem that the process of verifying the contents of the results of expert testimony submitted by expert witnesses is not properly carried out in practice. The parties are allowed to state their opinions on the results of expert testimony through dates for pleadings or fact inquiries in practice. However, the revision is intended to guarantee the process more clearly and procedural.

Expert witness examination can be divided into two cases: (1) the court requires expert witnesses to attend for the first time to take an oath, informs matters for expert testimony, and orders them to make expert testimony, and (2) expert witnesses make supplementary statements to the court after submitting the results of expert testimony in writing. In Article 339–2 of the Civil Procedure Act, expert witness examination can correspond with the latter. When the methods of examining expert witnesses were not separately stipulated, examining expert witnesses and the examination of expert witnesses under Article 340 of the Civil Procedure Act seemed not to be clearly distinguished in practice. However, those two are now clearly classified, and the examination of witnesses is defined differently by the revised regulation. In relation to examining expert witnesses, the method of witness examination by cross-examination does not apply mutatis mutandis to ease the burden on the parties. It follows the ex officio examination in principle, but the right to supplementary examination of the parties is guaranteed.

3.2.4 Examination of Expert Witness through Video or Other Transmission Systems

This regulation considers the burden on expert witnesses, such as the fact that they must argue with the parties and representatives face-to-face. And in case of they receive a small amount of fees because of the characteristic of public interest, the fact that they must appear in court as out-of-hours can be another burden. So, the regulation reflects that it is more reasonable to prepare the procedure wherein the court can hear statements without expert witnesses’ appearance in court. It would be better than to force them to serve summons. Unlike witness examination, which adopts the method of cross-examination, the expert witness examination follows the ex officio examination in principle. And it is also legally valid in relation to Article 339–2, which guarantees the right to the supplementary examination of the parties.

For expert witnesses, it is common to feel pressured by the questioning of the court and the parties about the results of the expert testimony. As a result, there is a situation where the court seems not to choose the expert witness examination to avoid the expert witness testimony [8]. Some pointed out that it eventually acts as the factor that hinders the acquisition of high-quality appraisal results and the effective end of disputes through it in practice. However, the court seemed to expect the effect of improving the phenomenon of avoiding expert testimony by allowing expert witnesses not to feel too much pressure during expert witness examination.

3.2.5 Entrustment for Expert Testimony

Entrustment for expert testimony means that a court may, if deemed necessary, entrust an expert testimony to a public agency, school, other organizations having adequate equipment, or a foreign public agency. The court may, also, have the person designated by a public agency, school, other organizations, or a foreign public agency make a supplementary explanation on a written expert testimony if the submitted statement is unclear or inconclusive (Article 341–2). The regulation applies mutatis mutandis to the examination of expert witnesses through video or other transmission systems during the examination of expert witnesses when the examination of the entrusted institutions for expert testimony is conducted. By this, it expects an improvement in the phenomenon of avoiding expert testimony of specialized institutions which can conduct entrusted expert testimony. In the case of the copyright appraisal under Article 119 of the Copyright Act, there was controversy over who should respond to an examination between an appraisal institution and an expert who actually conducted an appraisal. It was the case that showed the burden on the examination of expert witnesses. Also, it was expected that faithful and efficient evidence investigation could be conducted depending on the case.

3.3 System of the Procedures of Copyright Expert Testimony

Article 119 ① 1 of the Copyright Act allows the Korea Copyright Commission (KCC) to conduct an appraisal in which a court requests for an appraisal of infringement of copyright or other rights for a trial or investigation. In this case, the subject of the appraisal is the KCC. It conducts the appraisal in the manner of entrustment prescribed in Article 341 of the Civil Procedure Act, not as an appraiser as a natural person under the Civil Procedure Act.

Specific procedures or methods of the appraisal system under this Act are delegated to the Presidential Decree. Article 64 of the Enforcement Decree of the Copyright Act regulates the submission of the materials to deal with the duty. These materials are such as original copy or duplicate of a work subject for the appraisal, materials by which it can compare the similarity of related works in the case of a request for the appraisal on infringement. Other materials that the KCC requests as it judges necessary for the appraisal to the KCC and the organization of the Expert Appraisal Committee (EAC) could be the material too. Specific procedures of the appraisal system are delegated to the KCC. Moreover, under the principle of payment by the beneficiary, the KCC may request a copyright appraisal fee from a client, and the amount can be determined by the KCC. In civil cases, plaintiffs and defendants in the lawsuit are those who actually requested appraisals. So, they shall pay appraisal fees notified by the KCC to the competent court in advance. After appraisals are completed, the court pays the fees to the KCC. In criminal cases, appraisals are conducted after the state institutions, such as the court, prosecution, and police. So, they pay fees to the KCC, as the state is responsible for proving guilt [9].

Currently, the procedures of the copyright appraisal under the Copyright Act are divided into six stages: (1) application of appraisal and submission of appraisal materials (2) analysis of appraisal materials (3) estimate and notification of appraisal cost (4) appointment of professional appraiser (5) deliberation from EAC and (6) notification of appraisal result. In addition, additional materials may be requested if it is insufficient or inappropriate to conduct an appraisal after analyzing the submitted materials. Moreover, if it is necessary to revise the deliberation results of the EAC after the appraisal is conducted, it shall be corrected [10].

To request an appraisal, the court or investigation agency’s request for an official appraisal is required. Those who intend to request for appraisal shall submit applications of appraisal and materials to the KCC (Article 64 ① of the Enforcement Decree of the Copyright Act). It is necessary to analyze the appraisal at a certain level to calculate an appropriate appraisal fee and plan the appraisal following the request. Next, to conduct a full-fledged appraisal, the estimated appraisal fee calculated through an appraisal analysis must be notified, and the fee must be paid. In the case of a civil lawsuit, the person who actually requests an appraisal shall pay an appraisal fee notified by the KCC to the competent court in advance. In the case of a criminal lawsuit, an appraisal shall be conducted after an investigation institution pays a fee to the KCC. After the estimated appraisal fee is paid, a professional appraiser with the most appropriate professional ability conducts an appraisal. Multiple experts can be required to cooperate if necessary. An appraisal result is evaluated by the EAC, and the fairness, appropriateness, and reliability of the written appraisal are assessed so that they can provide a substantially and legally high-quality appraisal result. The appraisal result is then notified to the requesting institution, and follow-up actions such as the settlement of the appraisal fee, return of appraisal materials, and response to the questioning about the appraisal are made.

Such an appraisal procedure is sometimes problematic in that it somewhat conflicts with the court’s appraisal procedure under the Civil Procedure Act. A typical example is the procedural problems related to the provision of appraisal materials and the calculation of expected appraisal fees. The copyright appraisal under the Copyright Act proceeds with the procedures such as receiving materials and calculating estimated appraisal fees based on the entrustment for appraisal in practice. But the appraisal procedure of the court is to designate appraisers or entrust appraisals after confirming the estimated appraisal fees. In particular, it is very limited to provide materials for calculating estimated appraisal fees as appraisals have not been entrusted to the court yet. Moreover, there are certain problems including omitting important considerations for the selection of appraisers in practical appraisals under the Copyright Act. These problems arise as the court cannot confirm estimated appraisal fees before conducting appraisals based on the consent of the litigants.

Meanwhile, it is also worth considering that other institutions are not prohibited from conducting the same kinds of appraisals. Even though the appraisal procedure under the Copyright Act is institutionalized, it could be if the court may entrust the appraisals to them. The reason is that they seem to have similar practical issues in general.

3.4 Problems caused by using of these systems in intellectual property litigation

First, it is a matter regarding the difficulty of securing fairness. This represents a phenomenon in which the importance of the stage of appraisal in the trial process is linked to personal interests. A typical case for this is the construction appraisal corruption case in 2015. In March 2015, a builder and an owner filed a lawsuit over the cost of the new hotel construction in Geoje, Gyeongsangnam-do. In this lawsuit, the court appointed A as an appraiser, and he secretly offered both builders and builders to give an advantage. Knowing the importance of appraisal to the outcome of the trial, both parties each handed A KRW 10.8 million and KRW 8.5 million. In addition, B, the representative of the company to which A belongs, gave ‘subcontract’ to other technicians for the appraisal that A should take. Moreover, he submitted an appraisal under A’s name to the court. At this time, there is a mention that there is a practice that B usually receives about 10% of the appraisal cost as a fee. In particular, when an employee who left the company was selected as a litigation appraiser, B conducted an appraisal without the employee’s recognition and submitted it to the court [11]. This type of case was also the main cause of the revision of the appraisal regulations when the Civil Procedure Act was revised in 2016 [12].

The second problem is that the de facto binding phenomenon for appraisal results is increasing. As mentioned previously regarding the change in the type of appraisal, complex and highly professional litigation often occurs that judges cannot have various expertise corresponding to facts. So, a phenomenon in which judges are bound to the appraisal results may occur. In this regard, the Supreme Court prohibits judges from rejecting the appraisal results without clear grounds. The court said, “If the scientific evidence method proves that all the facts on the premise are true and the method of inference is scientifically justified, so the possibility of an error is regarded to be negligible or minimal since judges are quite binding in acknowledging facts, rejecting them without any reasonable grounds is prohibited as it is beyond the limits of the principle of free evaluation of evidence even if the fact acknowledgment is of the trial court’s full authority” [13]. In other words, it is acknowledged that the judge is bound to the appraisal result to a certain extent. If a judge has no choice but to be bound to the appraisal result beyond a certain level, the appraisal system is not simply constructed on the premise of assisting judges. But it must be reorganized to make the institutional system more advanced [14].

Third, it is a problem in accordance with the specificity of the use of appraisal procedures in the process of intellectual property trial. While the appraisal procedure is widely used in trials related to various professional fields, there are differences in using appraisal procedures. In intellectual property rights–related trials and other trials, the biggest difference is that the individuality of each case is considered in the trials related to intellectual property rights. When using the appraisal procedure in the trial process in general, an appraiser’s designation can be made based on certain rules. Replacing an appraiser with another expert also does not matter much. However, in trials in the field of intellectual property rights, factors such as expertise in the case and diversity of fields exist. Therefore, there is a big difference in that the process of investigating and designating appraisers suitable for solving each case should be dealt respectively. This characteristic leads to a problem in that the process of finding an appropriate appraiser. Also, selecting him/her as an appraiser imposes considerable difficulty and burden on the court in the course of individual cases would be a problem in trial practice.

Fourth, a problem is also pointed out that the closeness between appraisal results and judgment is very high. In reality, if the appraisal results are presented in a trial related to intellectual property rights, in reality, the appraisal results will inevitably have a very close impact on the trial’s final judgment. Accordingly, the court carefully addresses conducting the appraisal in the process of intellectual property rights–related trials. Unlike other fields, when an appraisal is performed in a trial related to intellectual property rights, it is common that the appraisal contents are basically connected to the matter of rights. For this reason, the reliability of the subject performing the appraisal plays a very important role. It is because that the judgment of the right to some extent is inevitably accompanied in some form in performing the appraisal [15]. Originally, judgments on facts and norms should be clearly distinguished into appraisers and judges. But these distinctions are not clear and somewhat mixed in the intellectual property rights–related trials. Thus, the reliability for appraisers has a higher demand compared to other fields. Furthermore, appraisals conducted by the appraisers with secured reliability — public confidence — are also important for the efficiency of the trial process as they can prevent unnecessary disputes during the trial.

Fifth, the high cost of an appraisal is also pointed out as a problem. In civil litigation, either or both parties in the case pay for the appraisal. However, even if it is a litigation appraisal, this point of cost becomes a problem in the case of criminal litigation. In the criminal procedure, the government bears the appraisal cost, inevitably resulting in considerable limitations to using the financial resources because it comes from the national budget. As a result, the high cost of appraisal becomes a burden, so it may be expected that courts will be unable to utilize necessary and appropriate appraisal. However, even in the case of civil litigation, the high appraisal cost can also lower the use of the appraisal. Moreover, since there cannot be a standard of appropriate appraisal cost, it is necessary to establish a series of standards of the appraisal cost calculated by reputable appraisers.

4. A method of determining the substantial similarity of a computer program using software forensics

4.1 The nature of the calculated similarity of a computer program

For a computer program work, quantitative similarity is used as a factual premise to determine whether there is substantial similarity.

Quantitative similarity refers to a numerical value (ratio) expressing the result of analyzing the source code of the two computer programs and comparing them by applying a certain standard. To calculate the quantitative similarity, the part expressed in the form of characters are the basic targets; for computer programs, the source code and the purpose code are the main targets [16]. This method shows the degree of similarity between the two computer programs, that is, the extent to which the infringer copied and used the copyright owner’s program in an objective figure. In addition, the presentation of such objective grounds is widely used as a basis to determine substantial similarity because of some advantages. For example, there are advantages of reducing unnecessary disputes in determining copyright infringement and deriving some generalized quantitative standards to promote legal stability. From a software engineering point of view, there can be various methods of calculating quantitative similarity. However, in order to determine practical similarity, it is necessary to establish standards and methods and calculate quantitative figures by reflecting the legal concepts and requirements contained therein. The quantitative similarity calculated without properly reflecting them is not appropriate for use as a basis for determining substantial similarity [17]. Therefore, in this case, it can be used as a basis for figuring out similarity, which is a requirement to prove whether it is based on the plaintiff’s material among the copyright infringement requirements. But it cannot be used as a basis for practical similarity.

4.2 Characteristics of quantitative similarity

First, it expresses the qualitative value of substantial similarity as the quantitative value by using numbers as a means of expression. It cannot be considered that the meaning of similarity is limited to the quantitative. It is because what is similar inevitably depends on the subjective intention of the person who determines it. In addition, the meaning of substantial itself is not drawn from specific quantitative standards as a qualitative value. But rather, it shows that the infringement of technology and effort through reproduction reaches the substantial degree to protect the value of technology and effort input by the creator to produce the work. It can be said that expressing these qualitative values in quantitative values through the numerical medium is a characteristic of calculating quantitative similarity.

Second, the calculation result of quantitative similarity is utilized as the basis for determining substantial similarity. Eventually, it not only plays a role of the basis for determining copyright infringement but is also used with expansion to the basis for stealing business secrets or calculating compensation for damages. In this case, it should be utilized with the recognition that there is a significant difference in concepts between the types of similarity for different purposes. It needs to be distinguished the similarity used as the basis for determining substantial similarity under the copyright law and the similarity used for other purposes. This is because, for example, if the similarity is used as the basis for stealing business secrets, it does not necessarily reach a significant degree but is meaningful for detecting that others use the content specified as business secrets. It is different from the area of copyright law, in which there should be something similar to a substantial extent.

Third, quantitative similarity represents similarity to both literal and nonliteral expressions. The literal expressions include an infringement on the character itself, that is, the literal expression, which appears on the source code line. Nonliteral expressions, on the other hand, include the structure, sequence, and organization of computer programs —source code [18]. Although they would all differ in extent depending on their functionality, they have a similar meaning to literary works. Also, the expression of similarity can be considered including both literal and nonliteral expressions [19]. In the case of nonliteral expressions, however, there are several controversies over various issues, such as to what extent copyright currently must be protected and how to calculate the quantified similarity. The stance of precedents also seems to be inconsistent.

Fourth, the value from the copyright law is reflected in the technical calculation method. This is due to the reality that the domain of judges intersects with that of experts in computer program copyright disputes that have the nature of professional litigation. It allows expert analysis and judgment to have an appropriate effect on the judgment. In practice, it seems that the protected scope of expression and the limitation of the concept of similarity are physically considered when quantitative similarity is calculated.

4.3 A method of calculating quantitative similarity

4.3.1 Overview

Looking at quantitative similarity from the perspective of technology, the methods for calculating a specific ratio can be widely varied. As mentioned, however, if this is not limited to the technical perspective but viewed as a value as close as possible to determine the substantial similarity under the copyright law conditions differs. There are restrictions in the calculation method arising. It can be seen that there are two major restrictions. First, similarity calculation should be conducted only by using the expressions protected by copyright law as a comparison target; second, the criteria for similarity determination should be established. This is because what is completely the same or different has no room for the intervention of subjective value. But similarity can yield significantly different results depending on what criteria determine whether they are similar or not. Of them, the latter case has relative difficulties. In practice, when comparing such as source codes, physical standards are applied to determine them as similar. These standards are mostly based on empirical rules. This is because specific consideration is insufficient on how to express the criteria for similarity determination as a physical standard.

4.3.2 Classification based on the expression method of similarity

Method 1 is a form that divides source codes and databases into file pairs to calculate the similarity of each and presents the results as they are. It can be used in cases where similarity is difficult to derive by applying one criterion because of different expression forms such as source code and database [20]. Method 2 is a form that derives the degree of similarity for each component (comparison element) and calculates and presents a single overall similarity by adding all of them. This method is mainly used to identify the clear degree of product unit infringement since a specific computer program is composed of a single product. While weighted value is sometimes applied to similarity calculation, the similarity is calculated by adding a certain weighted value in consideration of the significance of each component of a computer program [21]. However, some criticize that this is only a value calculated by a specific formula. They argue that it cannot be recognized as a value presented as close as possible to the actual similarity by calculating the weight problem or the comprehensive similarity forcefully.

4.3.3 Classification based on the calculation method of similarity

Method 3 is a way to calculate by comparing the degree of similar lines of the source code constituting a computer program with the entire line. It is to calculate the similarity faithfully to the literal expression of a computer program even if the functional part is slightly less considered [22]. Method 4 is a way that classifies computer programs by components (modules, files, etc.), determines whether each element is similar or not, and calculates the overall similarity by synthesizing them (Table 1) [23].

Classification based on the expression method of similarity		Classification based on the calculation method of similarity
Method 1	Method 2	Method 3	Method 4
Divide source codes and databases into file pairs to calculate the similarity of each and present the results as they are	Derive the degree of similarity for each component (comparison element);	Calculate by comparing the degree of similar lines of the source code constituting a computer program with the entire line	Classify computer programs by components (modules, files, etc.);
	Calculate and present a single overall similarity by adding all of them		Determine whether each element is similar or not;
			Calculate overall similarity by synthesizing them

Table 1.

Quantitative similarity calculation scheme type.

4.4 Limitations

First, it is the limitation of the calculated value close to the legal meaning. Upon examining precedents, it was commonly found that the basic concepts of copyright law are physically reflected for calculating quantitative similarity. Also, the two are compromised through comparison and analysis based on this. Nevertheless, no matter how precisely the quantitative similarity is calculated in consideration of the value of copyright law, there is a limitation in essence, that the actual similarity cannot be directly expressed. In this regard, the essence is the characteristic that the quantitative similarity is a technical analysis value. And thus, the process of its interpretation in terms of legal value must ensue.

Second, it is a limitation that the substantial meaning of quantitative similarity can be reduced. Changes on the computer program development environment are increasing the use of copyright non-protection targets. Therefore, there is a risk that the actual meaning of the quantitative similarity calculated may be distorted. It happens if this is excluded from the comparison according to the basic principles of the copyright law, compared and analyzed only with the remaining source codes. For example, if there is a source code of 100 lines and all of them were directly created, 60% of the quantitative similarity may be regarded as duplicating 60 lines out of 100. However, if 50 lines must be removed from the comparison target, such as automatic generation code and open source, the actual duplicated source code is 30 lines if there is the same 60% of similarity. In other words, it means that even though the quantitative similarity is the same, the actual amount of duplicated source code may vary. Therefore, even if both quantitative similarities are the same at 60% in a computer program with the same size of 100 lines, these two cases cannot be equally determined.

Third, it is a limitation caused by the distortion phenomenon because of the quantitative scale of the source code. The mistake that is most likely to be made in practice is to try to establish a quantitative standard for what percent of similarity is substantially similar. Quantitative similarity does not indicate the amount of source code judged to be similar in a computer program but simply presents its ratio. The actual amount of source code duplicated by the infringer cannot be identified only by the similarity expressed quantitatively. It is a matter that should be examined through a separate review.

Fourth, it is a limitation on the application of weighted value. Substantial similarity is the result of the determination of whether the quantity and quality of similar parts are substantial when comparing the comparison targets. Since the quality of similarity should be considered, the similarity may be calculated by reflecting the importance of the part that appears to be reproduced throughout the whole program. There is a case in which such qualitative importance is expressed as a weighted value and reflected in the process of similarity calculation in precedents and practices. However, applying the weighted value in comparing computer programs is controversial. It is because the question of how the appropriate the value assigned as a weight can be secured. In case law and practice, such qualitative importance is sometimes expressed as a weight and reflected in the similarity calculation process. However, it is somewhat questionable whether the weighted value can be applied so precisely in proportion to the difference brought by a minor change in figures. Finally, it is believed that applying the weighted value to consider the qualitative aspect in determining substantial similarity should be carefully approached [24].

Fifth, there are various methods of calculating the quantitative similarity of computer programs found in precedents. Despite the comparison under the same conditions, various results can be derived, because there are various methods of similarity calculation. This will act as the cause of a significant reduction of the predictability of each issue of copyright disputes over a computer program. In the end, a limitation arises that the inefficiency of dispute resolution is inevitably increased because of this. The Figure 1 is an example of deriving similarities when Method 3 and Method 4 are applied to two programs with a total of 500 lines and 5 comparison pairs. Although it is the same target, the result is 36.6% when Method 3 is applied, and 60.0% when Method 4 is applied, which shows a significant difference.

Figure 1.

Example of different calculated similarity by comparison method.

Sixth, it is a limitation of the difficulty in the clear establishment of the criteria for similarity. The establishment of the physical extent of the criteria for similarity are the most difficult problem in putting quantitative similarity close to substantial similarity. As discussed previously, the quantitative similarity of computer programs is basically calculated with text as a target. When comparing text, knowledge of how to set the criteria for similarity are very important as it affects the final degree of similarity.

5. Conclusions

Korean courts use the appraisal procedure under the litigation law. It is based on the ex officio principle to determine whether the copyright of computer programs is infringed. In this process, the quantitative similarity is derived to secure objective information on the degree of similarity, which is used as a basis for judgment of substantial similarity. Although several inherent limitations exist in this method, it is difficult to resolve them structurally. In practice, a method that secures reasonable methodology for each case is being used by fully considering these problems in the software forensics process. The result is shown differently depending on the method chosen to calculate the quantitative similarity. In other words, determining the method of calculating the similarity is critical in the trial outcome. However, depending entirely on experts’ goodwill is deemed to be inappropriate. Active discussions on not only a range of methodologies but also ways to improve the litigation procedures for the use of experts must be carried out.