We present a tool for the computation of the differences between two XML music score files and the intuitive visualization of those differences on the two scores side-to-side. Our goal is to implement for music scores a utility similar to the Unix diff command for text files. Its purpose is to identify differences between two score files which are relatively similar (typically two versions of the same file), corresponding to the intuitive notion of difference. The comparison is performed at the granularity level of bars (bars are here the analagous, for scores, of lines of text files), using a Longest Common Subsequence (LCS) algorithm, and, in a second step, at the granularity level of notes or chords (the analoguous of characters in text files). The later operation involves the computation of dedicated edit-distances between strings and trees, based on an abstract model of score content use for disambiguating the content for XML score formats and also decoupling our approach from the exact file format (in general MusicXML or MEI). The visualization is performed using Verovio and works for polyphonic scores and for multiple instruments. Straightforward applications are version control systems and collaborative music score edition, as well as speeding-up tasks that require human supervision, such as the visual inspection of the outcome of optical music recognition (OMR) systems.