Re: last archive slice

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Johnathan Burchill wrote:
| On July 25, 2004 01:10 pm, Denis Corbin wrote:
|
|>Johnathan Burchill wrote:
|>| On July 24, 2004 11:22 pm, Denis Corbin wrote:
|
|[...]
|
|>I would even say that you may miss a middle slice while dar knows
|>nothing that it is missing and keeps working properly restoring a subset
|>of files. Where is then disk archive as conceptualized object ? ;-)
|>(still quibbling).
|
|
| Is not a file an object? I can delete a character from the middle and it
| remains a valid file. An archive with missing slices may fail some CRC
| checks, but it is still an archive with (some accessible) data.

in the context of common human language, you are right, I think, a file
is an object one can manipulate. An archive is an object too.

|
| Actually, I am not convinced that archives should not be considered
| objects. For example, libdar does not give the user the option of
| specifying different backup directories for each slice.

dar could, but the only difficulty is how to have a simple specification
(compatible with command line restriction) that can assign a directory
to each slice number. Would the effort worth that feature ?


| The archive is in
| one directory. Libdar does not care whether that one directory is spread
| over different times, all the slices are still in that directory. When
the
| user is asked to provide a missing slice, it has to be placed in that
| directory.

the user can also have slice in many different directories and use
symbolic link that makes them appear in a single directory.

| As far as the user is concerned, an archive may not be a
| self-contained object, what with some of it sent over email, some on
| floppy, some on disk. But whenever libdar interacts with a given archive,
| all of the slices have to be available from the same source.

once an archive openned, dar still needs the good will of the user to
perform its action to the end. Required interaction may be far more
complex than requesting that the user put the correct tape in the
driver, if we want dar solve all the problem in place of the user. Dar
could scan the entire disk for the needed slice, keep a database of any
known burnt disks, seen floppies, and request the correct one in time,
etc... but that's not the direction I would like to take for libdar or
dar. I am not interested in making people more assisted, I am interested
in real feature a backup software can have.

|
| I must consider this further before coming to a conclusion.

me too. :-)

| [...]
|
|
|>That should be correct to my point of view. But you know some user
|>already told me that dar was too talkative, in particular about the
|>initial warning telling that all question will abort the program.
|
|
| A clever user. ;)

lol

who knows... ;-)


|>
|
|
| Okay, I will wait and see how things go in the API. Thanks for the
| discussion!

Thanks to you too. I may be a bit strict about the place of dar
relatively to the user (keeping things simples and rely on user's brain
for they understand what they do), but I stay open to discussion,
arguments against arguments, the best idea get the deal.

|
| JB
|
|

Denis.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFBBV3tpC5CI8gYGlIRAtWlAJ4+fIYGYSl+Ke2P9k/j6XAABEAi+QCfW8k5
BLYjceLIhuYdKgOHJP1XkTQ=
=1tRN
-----END PGP SIGNATURE-----



-------------------------------------------------------
This SF.Net email is sponsored by BEA Weblogic Workshop
FREE Java Enterprise J2EE developer tools!
Get your free copy of BEA WebLogic Workshop 8.1 today.
http://ads.osdn.com/?ad_id=4721&alloc_id=10040&op=click