Subject: Re: Long Filenames
From: Derek J. Balling (dballing@yahoo-inc.com)
Date: Thu Apr 06 2000 - 20:26:22 EDT
At 05:12 PM 4/6/00 -0700, Ron Chmara wrote:
><pedantic>
>Why this is wrong: The short name is actually the file *name on most
>95/98 disks*.
>Joliet tries to override the filename, by using the comment field,
>but, when given a choice between:
>A) Joliet filename, stored on disk and in files....
>OR
>B) 8.3 filename, stored on disk and in files....
>
>Your choices are often limited to what the actual _names_ are. That's why no
>program is inventing a shortened version if the filename is *already* short.
Well, that's PART of them, but the same problem/scenario exists where the
file is on the hard drive as well. I'll grant you that the "kludge" is
similar in both counts though. :)
>So the files are never renamed, but the transferring OS has to read what it
>_knows_ to be a filename. It doesn't invent anything, it just uses either
>the file name, or the file comment. What you are talkng about is adding
>a brand, spanking, new name, to replace the long comment. So then we're into
>many "names", with an actual ISO/DOS filename, one OS/2 file comment,
>(aka Joliet), and a _third_ name for MacOS 7-9, a fourth name for OS X
>over AFP, and ext2 name (if it breaks any of the above...)
ext2 actually would probably be the most forgiving of them all. ;-)
And remember that we're talking about transferring the file. So you're not
really "Adding" anything. You're replacing
/(8.3-filename|>32character_long_filename)/ with /MacCompliantFilename/,
and then, only on the destination host.
>Hence, the lowest common denominator, the name most OS's can handle,
>the 8.3.
Again, I'm not at all unclear as to WHY they did it, but (to my thinking),
if I'm looking at the file, under windows, I see a 40-character filename. I
don't see 8.3, I don't see anything else. That, to the user, is the
filename. (Semantics aside about how it is stored and you only get 40, 50,
80 characters through a monster kludge). If I transfer that somewhere, I
expect to have something... well.... moderately CLOSE to what it started
out as. ;-)
>Of course, many folks got so used to microsoft calling it a filename that
>folks now seem to think it _is_ a filename. :-)
Comments, by definition, wouldn't have to be unique. :)
Implementation-level aside, it IS a filename. It's a unique identifier to a
file. Whether that identifier is tied via a behind-the-scenes, one-to-one
hash with an antiquated 8.3 filename, an inode-id, or what-have-you, I (and
most people) view that as the filename. :)
D
-----------------------------------------------------------------------
Derek Balling 408-530-5062
Technical Yahoo Do You Yahoo!?
This archive was generated by hypermail 2b28 : Wed Jan 17 2001 - 14:30:25 EST