11 Mar
2011
11 Mar
'11
5:45 p.m.
From: Al Kossow
Sent: Friday, March 11, 2011 4:34 PM
On 3/11/11 4:03 PM, William Maddox wrote:
> I haven't had chance to look at it carefully,
but this looks like it
> is addressing the issue.
I had looked at this a while back. The problem is what
it is targeted
at solving:
"The original idea behind this project was to
provide a tool to apply
the data-recovery capability concepts of RAID-like systems to the
posting and recovery of multi-part archives on Usenet."
This doesn't map at all to use in a digital
repository, where you
(well, "I") don't want to explode a container and scatter it across
physical disks.
I guess the big philosophical issue I have with these
schemes is do
you wrap something around a data container (try to make it self-
describing), or leave it as two separate containers ("read-only" and
"up datable").
Where do you come down on that question?
My personal preference would be for something of a meta-filesystem,
similar to the Mac OS X packaging for applications (a philosophical
outgrowth of the resource/data forks idea simplified by using a file
system to implement it). One fork is the actual archived bits in
whatever form best preserves it--tar balls remain tar balls, tapes
become streams of octets, etc.--while another is an analogue of the
TCFS vel sim. as a metadata description of the content of the bits
in the archive. The tools to get at each piece are simple, but the
packaging tools allow the archive to be treated as a unit.
"But that's just my opinion, I could be wrong." --Dennis Miller
Rich Alderson
Vintage Computing Sr. Server Engineer
Vulcan, Inc.
505 5th Avenue S, Suite 900
Seattle, WA 98104
mailto:RichA at vulcan.com
mailto:RichA at LivingComputerMuseum.org
http://www.LivingComputerMuseum.org/