On Mon, Oct 3, 2011 at 8:37 AM, Stephan Aßmus <superstippi@xxxxxx> wrote: > I am not sure what exactly the purpose of file is, but the expectations of a > CLI user may depend on the exact situation. For example, bypassing the Haiku > mime detection may actually be the point of using file. The file command inspects the data in the file, then tries to match it to a known pattern. Many well-known file formats start with a "Magic Number" which is quite often 32 bits. For example TIFF files have the ASCII either MM or II for a byte order marker (Big or Little, Motorola or Intel) followed by the ASCII for TIFF (or maybe the byte order comes after the magic number, I don't recall). On Mac OS X the file command can tell if executables and libraries are Universal, that is, whether they have 32-bit or 64-bit code or both, as well as containing PowerPC or Intel or both. A problematic case is OpenOffice.org files, which are Zip archives of a small directory heirarchy full of XML files. The problem here is that the compression messes up the recognizable data bytes. I would think that many scripts - or users - would want to set the mimetype based on what the file command reports, rather than the other way round, that is, having file check BEOS:TYPE. -- Don Quixote de la Mancha Dulcinea Technologies Corporation Software of Elegance and Beauty http://www.dulcineatech.com quixote@xxxxxxxxxxxxxxxx