"File types"

[Don]

Chuck Guzis wrote:
> On 8/29/2006 at 12:28 PM Don wrote:
> 
>> I.e. type needs to be finer grained than silly N *character*
>> file "extensions"... *everyone* wants to be "DOC", or some
>> other pronounceable extension.  So, the file types lose their
>> value (e.g., on my W2K box, Matlab and Mathematica both want
>> to use .m -- so, .m files are meaningless to me since I can't
>> recall FROM THE NAME OF THE FILE which application needs to
>> be opened to process the file).  MacOS *seems* to have had the
>> right idea (though I have never used it "seriously" to know
>> for sure).
> 
> ...and you might add that there are no rules for "extensions".  In front
> of me, I've got a disk with .FW and .PCL files.  And no--they don't have
> anything to do with Frameworks or HP.  Neither is listed in Filext.com.

Exactly.  I have .DAT files that 3 different applications want
to claim as their own.  About the only "file type" that I
have NOT had problems with is ".dgy"  :>

> Since Windows closed the barn door after the horses ran off, so to speak,
> the use of extension to identify the application associated with a file is
> not particularly wonderful.   I don't know if it could be dangerous.

It adds no value, IMO.  And, only increases the chance that
the *wrong*/unexpected application can be invoked without
your being aware of it beforehand (I can list dozens of
"extensions" on my W2K box that are ambiguous -- several of
which invoke the *wrong* application handler.  Hopefully,
each of these handlers are smart enough to parse the file
given to them to determine if it, in fact *is* a valid
<XXXX> file before operating on it.

So, the type is merely a convenience for The Desktop.  And,
IMO, a poorly designed/implemented mechanism.  :-(

(I *really* need to play with MacOS more and see what the
"shortcomings" of their scheme is/was -- aside from file
portability)