17 May
2005
17 May
'05
5:18 p.m.
On Tue, 2005-05-17 at 16:43 -0700, Vintage Computer Festival wrote:
On Tue, 17 May 2005, Al Kossow wrote:
Last random thought before bed...
Does it make sense to split the spec to handle this?
One that dictates how the metadata and raw data are stored that make up
an image, and another that then defines how to wrap all of this in a
layer that's properly checksummed etc.?
Blah... I know I'm going to have a bazillion messages to read by
morning!
cheers
Jules
You do need to be mindful of purging copies of
bad files of which there
are other known good copies, and you need to make sure that the verified
files have some error detection/correction mechanism to detect silient
copy corruption (failures in disc write caches, etc.)
A feature design into the draft FK spec.