"Russell L. Harris" <russ...@rlharris.org> writes: > = With git, I can do a hundred backups in a morning or afternoon > session without worry about consumption of disk space. I am not > concerned with an orderly and uncluttered commit record; my concern is > the ability to recover.
Git is an awesome tool that can be utilized with a wide variety of files, not just software source code or text. However, it faces scalability challenges with larger files, such as videos. While extensions like git-annex can assist in managing these larger files, they can struggle when dealing with tens of thousands or more of files. Additionally, Git's core feature of maintaining immutable history is not well-suited for backing up large, dynamic datasets, where retaining the entire history indefinitely may be undesirable. There may also be privacy and legal concerns related to maintaining permanent records of all changes, since establishing a retention policy is difficult. So, most people cant't have their whole $HOME in a Git repository. Moreover, storing Git repositories on the same storage device as your original files only safeguards your data against your own mistakes. This is analogous to the (opposite) misconception that RAID systems serve as a backup solution; they only protect against hardware failures of one or more disks, and happily mirror all your mistakes for you. Proper backup system need to protect your data from most (ideally all) data loss scenarios. Fortunately, with Git, setting up remote copies on another disk or machine is straightforward, so it is an excellent backup tool for data sets that are well suited to version control, such as your writing projects. > = The time required for each backup is negligible: > > (1) switch to a virtual terminal > (2) type: git commit -a -m journal #050 > (3) switch back to Emacs I recommend checking out the Magit package for Emacs. Once people try it, few return to using the command line for git operations.
