Right now the list of extractors to which a file can be submitted to in the "Submit for extraction" page is not filtered by compatible mime types. This means that a user can submit a file to an extractor that might not be compatible. For example a PDF file to the ncsa.nlp.simplesummary extractor.

Even though the flexibility to try different extractors in case the work is nice, we should respect the mime types defined by the extractor and only make those available.