Looking at "bit-rot", and the life of your data

[GreyGeek]

Linux Magazine - Checksumming Files to Find Bit-Rot

[woodsmoke]

Re: Looking at "bit-rot", and the life of your data

Hi
That was a very interesting article.

So, I might guess that the methods outlined in the article to minimize or eliminate "bit rot" would be a good thing to be implemented in the next iteration of Kubu file system, or possibly the "upstream" Ubu file system, if they have not already been so implemented.

However, the concept raises several points.

a) if it has already been done and it was in one of the release notes: "We fixed/changed yada in the file system".

But, the devs are disappointed at the response to all their work because of the basic format of how a release is announced and all that work was lost on the user.

A "reverse psychology" to that is one of the older dictums of manufacturing. If one "talks" about a process then others "know" about the process and then it is a continual spiral of "look at how I improved things compared to those guys" and it again, comes down to cost/benefit, what is more important? something that is flashy and gets new users/maintains base, as opposed to something that is REALLY of benefit, but is massively labor intensive and is never seen by the user because the emergency of running screaming into the night doesn't happen.

Another take on it is from science. It is impossible to prove a negative. It is hard to show that when something does not happen that that is a good thing.

b) if it has not already been done and is being contemplated what will be the cost/benefit ratio in terms of the amount of work, cd space, you name it compared again.....to the response by the user as evidenced at Distrowatch?

c) If there is not a cost benefit ratio that is acceptable in terms of development or the delivery system, is there a way that some kind of "file watcher" could be placed into the repos, which would install every time, all the time, so that the user at least "thinks" they have done something and that would actually "do" something better than the regular file system in a way that the user can "see" that their information is being kept.

d) Is any of this improved or degraded by "static" storage, my term, in terms of say a "USB stick" as opposed to a rotating drive?

e) Again, in terms of cost/ benefit, is it just hands down better to move information to a plastic medium like a cd (which, unfortunately, are being phased out) and let it go at that?

just some thoughts which may or may not be easily answerable.

woodsmoke