Hacker News Reader

Comment by mort96 | Hacker News Reader

[flagged]

Zstd doesn't use a login, it doesn't track you across the internet, it doesn't make shadow profiles on you, and it doesn't have ads or telemetry. Zstd is even part of the Linux kernel at this point[0], so it's safe to say that it isn't a Facebook product anymore. They can't pull the rug on zstd, since it's BSD/GPL licensed. Unless I'm missing something, nobody is feeding the beast by using zstd.

There's lots of similar cool technologies made by other "evil" companies. I don't use ZFS, but its origins in Sun/Oracle isn't the reason I don't.

[0] https://lore.kernel.org/lkml/20181109190304.8573-1-kilobyte@...

I don't use ZFS, and the Oracle origins is the reason why. I'm very worried that the best filesystem available is forever stuck with a hostile license, and that so many of the good filesystem engineers are contributing to Oracle's product rather than to a non-Oracle alternative. (EDIT: I know zstd is under a good license, so its situation isn't the same a ZFS's, I just commented on ZFS since you brought it up.)

Anyways, I prefer to not use Facebook products.

> that so many of the good filesystem engineers are contributing to Oracle's product rather than to a non-Oracle alternative.

OpenZFS isn't Oracle's product; it certainly includes a great deal of Sun/Oracle code, but Oracle themselves can't use the OpenZFS commits without honoring the CDDL license, which to my knowledge they've not done.

Fair, but all the Oracle code in OpenZFS precludes OpenZFS from ever re-licensing under a less hostile license.

ZFS was made by Sun Microsystems, which was a dear, awesome company that educated, fed, clothed and put bread on the table for an entire generation of computer professionals outside of Sun Microsystems, not to mention paid millions out of their own coffers to open source a reference implementation of a System V Release 4.0 UNIX. And to top it all off, Sun Microsystems made awesome hardware.

  Sun kicked butt
  had fun
  didn't cheat
  loved their customers
  changed computing forever.

I hear people say that, yet Sun Microsystems is the company which created the CDDL to be intentionally GPL-incompatible[1] and then made a filesystem which they licensed under the CDDL. Then they sold themselves to one of the worst tech companies. Not exactly the behaviour I would expect for an awesome company.

[1]: https://fedoraproject.org/wiki/Licensing/CDDL, https://en.wikipedia.org/wiki/Common_Development_and_Distrib...

I have some internal conflict about zstd. On the one hand, it's truly become the superior general purpose compression codec, and I hate to look a gift horse in the mouth there. Using it for everything from data-at-rest compression on filesystems and zswap to channel compression with rsync, it truly works wonders. And yet, I can't really deny how absolutely at odds I am with the business model and actions of the company that has sponsored its development. Personally, I've resolved that we should extract as much good from things I perceive as evil. I've stopped using Facebook platforms myself, so I'm doing as much as I can to reduce the amount of money I generate for the platform as I can. That's about the most "voting with my dollars" I can do. However, because zstd has been such a net good, I will continue to use it in projects going forward for compression needs.

I'm a Facebook hater too, but I can't think of any even theoretical concerns around zstd. It's freely licensed, no patent concerns, it's straight up free software. You're not giving them money or supporting their ad model by using it. It's not like Chrome where Google has an interest in making web browsing suck.

What theoretical path to harm is there to adopting it more widely?

Petty perhaps, but I'm really annoyed that he didn't follow the celebrated command line tool tradition of compress, gzip, xz, lzip, bzip2, etc. I always have to remember that zstd behaves differently and you have to take extra steps to achieve the same thing.

What differences do you have in mind? Zstd is largely compatible as a drop-in replacement for those commands. There are differences, but it's our goal to minimize them.

Did you try it? compress/gzip/bzip2/xz deletes the source file whereas zstd leaves it behind. This applies both to compression and uncompression.

I think the best way I can describe my discomfort is this: If someone asked you: "What organization do you want to be the owner of, and own the copyright to, the compression library which underpins all our open standards and systems?", what would you have responded? I don't know for sure which organization I would've chosen, but Facebook would certainly be at the very bottom of the list. GNU and Mozilla might be closer to the top.

Weird take seeing as it's licensed under GPL.

As the copyright holder Meta, its affiliates (e.g., Facebook) or their successors, can change the license for zstd at any time.

It is a BSD license, not GPL.

https://raw.githubusercontent.com/facebook/zstd/dev/LICENSE

> As the copyright holder Facebook, or its successors, can change the license for zstd at any time.

This is wrong? Once something is GPL, it's always GPL. https://softwareengineering.stackexchange.com/a/98777/347066

It's a bit more complicated than that. Yes, the code that's out right now under the GPL will always be licensed under the GPL. But Facebook can at any time change the license of the zstd project, meaning all future releases will be under a different license. There will be an opportunity to fork the code base pre license change, but those forks are always risky and there's a very large change that the ecosystem won't end up with a de facto standard zstd fork.

Reply on news.ycombinator.com