[londons_explore]

I'd like to see a version of this built into things like IPFS.

It seems obvious that whenever something is saved into IPFS, there might be a similar object already stored. If there is, go make a diff, and only store the diff.

[hcs]

It should be possible to do this in IPFS already if you use the go-ipfs --chunker option with a content-sensitive chunking algorithm like rabin or buzhash [1]. With this there's a good chance that a file with small changes from something already on IPFS will have some chunks that hash identically, so they'll be shared.

[1] https://en.wikipedia.org/wiki/Rolling_hash#Content-based_sli...

[londons_explore]

But that isn't quite as good as something like this that can 'understand' diffs in files, rather than simply relying on the fact a bunch of bytes in a row might be the same.

[hcs]

I don't think elfshaker actually does do any binary diffing (e.g. xdelta or bsdiff). It works well because it uses pre-link objects which are built to change as little as possible between versions. Then when it compresses similar files together in a pack, Zstandard can recognize the trivial repeats.

[peterwaller-arm]

Author here. This is correct, we set out to do binary diffing but we soon discovered that if you put similar enough object files together in a stream, and then compress the stream, zstandard does a fantastic job at compressing and decompressing quickly with a high compression ratio. The existing binary diffing tools can produce small patches, but they are relatively expensive both to compute the delta and to apply the patches.