RE: Naming/Format conventions for INDEX files

[Markus Stumpf <[e-mail]>]

Hoi!

|>>MX_ROOT:[MLF]00README.TXT;1                        13   6-SEP-1991 08:09:54.96
|>>MX_ROOT:[MLF]BNB.P;1                                6  29-JUL-1991 16:49:02.68
|>>
|>>The format is:
|>>disk:[directory]filename.extension;revision     blocks date        time
|>>                                  (1 block = 512 bytes)

|>
|>How about the VM/CMS format or a CDC NOS/VE format? :-)  Seriously,
|>if you are going to propose a standard, don't make in Unix-centric.

Why don't we collect a list of the formats the FTP programs of the various
OSs support/need to retrieve files?

For the various OSs formats I think that the format proposed by me should be
split in two parts: The size and date part and the file naming part!

The size and date parts are NOT Unix standard! They simply reflect data
retrieved from the OS! This should be possible on any OS (as we see it IS
for VMS and I know it is for DOS and VM/SP CP/CMS). The format in which
it gets display should be (I think) as simple as possible!
I chose the yyyy-mm-dd hh:mm format cause it is better readable than a
yyyymmdd hh:mm or yyyy mm dd hh mm and easier to sort than using (abbreviated)
month names.

For the filenames I think we should use some naming convention that could
be used with most (better ALL) FTP-Servers!

All the MSDOS FTP-Servers I know about support the Unix naming conventions!
As for VM/SP CP/CMS - as I can remember (it's about one year we gave our
mainframe away) - one is able to retrieve a file from minidisk 191 that is
called bla.bla.a with the command
	'cd 191'
	'get bla.bla.a'
(correct me please, if I'm wrong).
I'd like to know which commands to use for the VMS/FTP, to retrieve e.g. the
above mentioned file MX_ROOT:[MLF]00README.TXT;1!

To be clear:

We should have a format for the index files, which can be used by FTP-Servers
and not the format which is the output of the 'ls', 'flist', 'listfile' or
any other command that provids an OSs file-listing capability. We should also
keep in mind to have it simple and be supportable by all OSs.

For me we could even use something like

size   date time   disk':['directory']'
size   date time   disk':['directory']'file
size   date time   ':['directory']'file           (if no "disk" info is avail.
                                                   or needed)

or some other characters or some other indication to seperate the information
directory and file.

|>>The earlier you think about something other than Unix, the better!  Right
|>>now, worrying about format is trivial -- worrying about the interface
|>>between OS'es (rather than flavors of the same OS) is what needs to be
|>>looked at.

Right! We need a generic file naming format, which supports all platforms
and FTP-Server programs!
To use this information to actually retrieve a file is up to the user or
the wrapper around the "ftp" programm.
But we should really try to keep this format as simple as possible!

|>>            To me, this is not trivial as I will very soon be working on
|>>the administrative team for parallel Unix and VMS AFA hosts out of the same
|>>site -- which I would like to see somewhat parallel.  The IAFA, at least in
|>>title, gives me hope; in practice, though, I am in despair.
|>
|>I fully agree.  Lets not make this into the Unix IAFA.

Unix IAFA ???
^^^^
Why Unix?? And I think this is an really important topic! What does an
ftp archive give to me (oh, my bad english) that has a lot of good software,
but which makes it nearly impossible for me to retrieve it in an acceptable
amout of time!
And the amount of time needed, is in my experience, strongly dependant from
the quality of the index file! And a standardized format of this file could
make this time decreasing!

Ciao
	\Maex
-- 
______________________________________________________________________________
 Markus Stumpf                        [e-mail]