A GSG CASE|
Camco vs. SOFT, Inc.
Camco and Soft, Inc. are fictitious names chosen to disguise the identities of the parties in this case history from the files of General Systems Group.
General Systems Group (GSG) was retained by Camco to compare a complex computer aided design/computer aided manufacturing (CAD/CAM) software package Soft, Inc. had developed for Camco to a package that Soft, Inc. had subsequently developed to market on its own behalf. Camco alleged that Soft's later package was suspiciously similar to the package developed for Camco's exclusive proprietorship.
The Camco owned package consisted of 69,000 lines of C source code, or approximately 1,800,000 bytes of source code. The system included 150 distinct source modules, (separate programs for different tasks) and over 800 subroutines.
The package that Soft, Inc. developed to market under its own name consisted of about 31,000 lines of C source code or approximately 920,000 bytes of source code. The system includes 31 distinct source modules and a little under 600 subroutines.
Camco engaged GSG to determine whether there was a basis for suing Soft, Inc. for appropriating the intellectual property represented by the package that Soft, Inc. had developed originally for Camco. Camco alleged that the two packages in question performed similar functions and were marketed to the same audience. Camco assumed since the same Soft, Inc. designers had worked on both packages, and since the development time for the second package was quite brief, that Soft, Inc's own CAD/CAM package had to be the result of extensive plagiarizing of the source code of the package Soft had developed for Camco. So GSG was hired to compare the two sets of source code and determine if any plagiarism was involved.
We proposed a two phase approach, namely: a short, inexpensive phase of a couple of weeks effort, to determine the degree of relatedness, if any, of the two sets of sources, and an optional, more expensive follow-up to document in depth the actual relation between the two sources. The rationale for this approach is that software plagiarism can be nonexistent, blatant (crude), or very subtle. In the former two cases, a brief analysis is sufficient to obtain the necessary evidence, and therefore to rule out the third case. If the third case cannot be ruled out, the short examination can give evidence as to the nature of the subtle form of plagiarism involved, and therefore an indication of the level of effort (cost) required to prove it.
What GSG Found
Camco agreed to the two phase approach, and our preliminary analysis found that the two packages were unrelated and that there was no evidence of plagiarism. This finding profoundly reoriented the Camco legal team's litigation strategy. Rather than alleging plagiarism, they would allege violation of the noncompetition clause of the development contract between Camco and Soft, Inc.
The specific way we established this finding was to prove that all low level "libraries" of subroutines for both packages, while respecting identical specifications, were coded in substantially different ways, and for a few critically selected very high level functions, the kind of functions that embody the application know-how, that both the specifications and implementations were different.
This approach is based on the fact that the most likely body of code
to be plagiarized by a designer, in a subsequent design, is the set of
libraries of low level subroutines, because they are essentially his
building blocks. The next most likely set of code to be plagiarized
is high level functions which contain a high degree of the
"application smarts" because they contain the most complex and expensive
code. In sophisticated plagiarism cases functions between these two
extremes are much less likely to be the target of intellectual