Did anyone mention MacBinary, the sanctioned method for wrapping
two-fork Mac OS files in a single file for transmission or
storage in a single flat file? Or the way some apps were
smart enough to create Mac files as MacBinary when the files
were being stored on non-Mac (networked) file systems?
Or AppleDouble format, the way they preserved the fork on
non-Mac floppies (both MS-DOS and ProDOS)?
Did anyone mention the early NT Server's optional Mac compatibility
mode that allowed Macs to store forked files on NTFS?
Did anyone mention the way some old Mac apps used the resource fork
(not the data fork) for storing data, because there were
convenient API calls that made it slightly database-like?
Did anyone mention the way Windows hides the extension from the
user by default?
Would it be useful to discuss old OS filesystem metadata, such
as the 80-char "comment" field in AmigaDOS?
No one mentioned the
http://www.formatexchange.com/ project
from a year ago. Sellam was promoting it.
As for file.dmg.hqx and file.tar.gz, did anyone mention the Icelandic
naming habit of assigning a child's last name based on the same-gender
parent's first name? Annasd?ttir? ?orvaldsd?ttir? Gunnarsson?
At 06:10 PM 8/29/2006, Tony Duell wrote:
To return to the original question, how far back do you
have to go for
unix's cc(1) to require .c on C source files, .o on object files (to be
linked it), and so on?
I wonder if there exists a timeline of when certain Unix concepts
came into being. 'file' and magic, for example. To me, it shows
that it can be valuable to have an automated method of knowing
what's in a file, even for believers in streams of bytes.
Tomorrow's OSes will be far more like databases of metadata, I think,
than inodes for streams of bytes. Easily recognized, preserved and
parsed metadata makes all sorts of interesting applications possible.
Geotagging. RFID. GUIDs for everything. EXIF in images can be
very handy.
Yes, it's easy to think of twists and permutations that will make it
complex or problematic to preserve metadata as files change and mutate,
but I'm sure people will continue to invent new apps to justify it.
I've always thought that even having just "genus" and "species"
within the metadata for files would go a long way. Three-char
file extensions don't tell much. Apple's old notions of
type and creator were handy, too.
- John