On Wed, 2007-04-18 at 21:43 +0200, Gérard Milmeister wrote: > Looking inside a file to determine a mime type is just another hack. > The only correct way to do it is to attach meta-data to each file > (using extended attributes, for example). However this would mean > quite a radical change of how files are created, modified etc. I think meta data should be in the file, not a separate resource (and that means *IN* the file, not seeming like its in the file because the filing system is amalgamating its meta data with the file contents, on the fly). It gives you easy tranmission of the file. And if the meta data is done sensibly, quick assessment of what the file is. Of course, as my prior message outlined, it'd have to be done in a uniform way in the first few bytes of the file. Otherwise you will bog down a system trying to represent the contents of a large directory of files. The Amiga IFF system was a fairly good example. You had descriptive header information in a file, and content, each programatically defined in a standard manner. -- (This box runs FC6, my others run FC4 & FC5, in case that's important to the thread.) Don't send private replies to my address, the mailbox is ignored. I read messages from the public lists.
