Re: "in-program" backup facility
- To: mathgroup at smc.vnet.net
- Subject: [mg111873] Re: "in-program" backup facility
- From: Bill Rowe <readnews at sbcglobal.net>
- Date: Mon, 16 Aug 2010 05:55:22 -0400 (EDT)
On 8/15/10 at 7:36 AM, plindsay.0 at gmail.com (peter) wrote: >I've seen a few posts about this and I'm wondering if I am missing >the point [ again ]. Surely folks are managing their own backups - >perhaps using "Time Machine" on the mac or whatever. The arguments >in favour or regular backups surely extend beyond the use of >Mathematica and apply generally to the use of a personal computer ? Undoubtedly, such backups are done by many users here. But, such backups don't entirely resolve the issue. Time Machine and the like backup the entire hard drive. The default interval for Time Machine is 1 hour. With Mathematica, a substantial amount of work can be lost when restoring a notebook to what it was 1 hour ago. And note, Time Machine backs up the file as it exists on the hard drive. If you have not saved your work periodically, the file backed up by Time Machine could be quite a bit more than 1 hour old. Creating some sort of script to automate periodically saving the state of a notebook and decreasing the interval Time Machine uses for backups isn't a viable solution either. If you decrease the interval between backups enough, then Time Machine will be perpetually running and never complete a backup. Further, with a scheme to automatically save notebooks there will be the chance Time Machine will be writing the notebook to the backup at the same time it is being modified (saved) by the automated save script. And the probability of this occurring clearly increases as the interval between backups is decreased. The result of such collisions will be the copy of the notebook in the Time Machine backup won't be stable and likely is unusable. Programs designed to backup all files that have changed on a hard drive really aren't replacements for what is being asked for.