Railroad modelers strive for historical accuracy. However they will vary a model from an existing building in two respects. The modeler will practice ‘selective compression‘, which involves removing details in order to fit the smaller scale, while retaining the look of the original. Second, they may vary the appearance of ‘trim’: color and design of, for example, window frames, awnings, etc.

Plaintiff sues competitor/defendant for copyright infringement in three models. Defendant’s model has 20 or so similarities with plaintiff and 20 or so differences (however the decision doesn’t provide great guidance on how significant these differences or similarities are). Defendant alleges that plaintiff’s works were based on public domain building (and didn’t disclose as such on its copyright registration application).

Defendant’s summary judgment is denied. Decision discusses (1) what constitutes fraud on the Copyright Office; and (2) what should test of similarity be for ‘thin’ copyright. Defendant’s proffered ‘super-substantial similarity’ is rejected in favor of normal test merely applied to fewer elements.

43(b)log presents a comprehensive discussion of the issues here.

Decision Osment Models v Mikes