Skip to Main Content
Computing differences (diffs) and merging different versions is well-known for text files, but for models it is a very young field - especially patches for models are still matter of research. Text-based and model-based diffs have different starting points because the semantics of their structure is fundamentally different. This paper reports on our ongoing work on model-independent diffs, i.e. a diff that does not directly refer to the models it was created from. Based on that, we present an idea of how the diff could be generalized, e.g. many atomic diffs are merged to a new, generalized diff. One use of these concepts could be a patch for models as it already exists for text files. The advantage of such a generalized diff compared to dasianormalpsila diffs is that it is applicable to a higher variety of models.