I understand that git lfs replaces the lfs tracked files with pointer files in my repository.
What I don't understand is what I gain from that happening.
Is it just that I never have to download the binary files associated with commits that I never checkout?
Or are there other advantages?
It boils down to the reduction of clone time, storage space - for those who clone the repository - and network bandwidth use for both the client and repository server.
It may be just that for the average user and some even see it as big disadvantage since lfs introduces centralization, at least of parts of the repository.
But especially for companies and git services that have many users and/or many versions of big binary files, this can be a great saving of resources like storage, bandwidth, cpu time and so on. These are normally very centralized anyway and could f.e. throttle usage of the LFS server while maintaining the "normal service".
The Roadmap includes a bit more, like metrics, resumable downloads which would be new and maybe advantageous to the normal git with big files. The rest is more or less an improvement for lfs.
