For those not knowing what this is about
Disclaimer: I am no expert on this subject. This is just a general overview of different sources. Please correct me on anything that needs correction.
With pacman v5.2 Arch Linux has officially included support for a compression format developed by Facebook, Inc. called zstd. You can read an early proposal from March this year on the arch dev mailing list here and check the Github Repo. Arch is also planning to convert this to their standard compression method for packaging where formerly they had used xz instead which is based on LZMA.
This format is actively being developed by Yann Collet and Przemysław Skibiński at Facebook Inc. since February 2015. It is a multi threaded compression format in contrast to xz being single threaded. Arch developers have apparently decided that this would make for a way better compression ratio based on benchmarks which can be found especially in the mentioned mailing list.
From our end, this whole topic came up because of this thread.
Why this may raise concern in some people
Facebook is one of the so called "Big Four" companies and without a doubt one of the most powerful companies of this society we live in around the globe especially when it comes to private data. This is basically because we willingly tell Facebook everything about ourselves, but there's also a lot of code included in Facebook based services and servers that will try to get even more juice out of your not so obviously shared information. Basically things you never opted in to share. Like your actual physical location at any given moment, your telephone number, your interests and how much porn you really watch. Often this is data that can be collected because you use some service or another that is directly connected or owned by Facebook and you did not actively opt out of harvesting options activated by default. This is true not only for Facebook but for a lot of companies that make business with data, which are quite frankly the vast majority of all services you use on the internet.
Such companies are apparently also of high interest to governmental institutions such as so called intelligence agencies which are part of the military apparatus of a country which in return can usually provide vast amounts of money being (usually) the countries major investment. This is no news in itself and data harvesting companies in several countries (I suppose most) are even obligated to hand out information if it would assist in an investigation.
What's this gotta do with a compression format!?
Coming from one of those major companies the compression format in question is written in C and published under the BSD License and GPLv2 on Github. Connecting this with the former paragraph it is rather apparent that this might empower this specific company to become more and more a major part of the Linux community. Linux has grown and matured a lot and even gained user base on personal computers and it appears companies are getting aware of it. Also see Fungalnet's article on his blog about this where he mentions the recent purchases of Github and RedHat.
Being a part of something also means a certain influence and possibilities to exploit that something. This is true for our very life and any social construct to begin with. In this situation it may open way for Facebook to inject questionable code to harvest data into Linux. This is basically what this is all about.
My very personal opinion
Disclaimer: This is my personal opinion based on the little information I have on this matter. Please correct me where correction is due. Also I'm too lazy to search for sources here, as most of this is just very basic stuff.
To me personally the reason behind putting money and effort into the development of a multi threaded compression algorithm would suggest interest in pushing faster mobile and embedded device operation and gaining market share by being an active force in said development (see concerns and what that's gotta do with anything). AMD and Intel are fighting heavily for market share on PC/professional with immensely powerful work station and gaming CPUs getting smaller and smaller while ARM is already dominating the mobile market with just as powerful and more energy efficient processors for micro environments, which in reality is way bigger nowadays anyways.
For me there's several sides to this:
Discuss, share your views, your knowledge, your opinion!
( In a civilized manner apparently ;) )
Disclaimer: I am no expert on this subject. This is just a general overview of different sources. Please correct me on anything that needs correction.
With pacman v5.2 Arch Linux has officially included support for a compression format developed by Facebook, Inc. called zstd. You can read an early proposal from March this year on the arch dev mailing list here and check the Github Repo. Arch is also planning to convert this to their standard compression method for packaging where formerly they had used xz instead which is based on LZMA.
This format is actively being developed by Yann Collet and Przemysław Skibiński at Facebook Inc. since February 2015. It is a multi threaded compression format in contrast to xz being single threaded. Arch developers have apparently decided that this would make for a way better compression ratio based on benchmarks which can be found especially in the mentioned mailing list.
From our end, this whole topic came up because of this thread.
Why this may raise concern in some people
Facebook is one of the so called "Big Four" companies and without a doubt one of the most powerful companies of this society we live in around the globe especially when it comes to private data. This is basically because we willingly tell Facebook everything about ourselves, but there's also a lot of code included in Facebook based services and servers that will try to get even more juice out of your not so obviously shared information. Basically things you never opted in to share. Like your actual physical location at any given moment, your telephone number, your interests and how much porn you really watch. Often this is data that can be collected because you use some service or another that is directly connected or owned by Facebook and you did not actively opt out of harvesting options activated by default. This is true not only for Facebook but for a lot of companies that make business with data, which are quite frankly the vast majority of all services you use on the internet.
Such companies are apparently also of high interest to governmental institutions such as so called intelligence agencies which are part of the military apparatus of a country which in return can usually provide vast amounts of money being (usually) the countries major investment. This is no news in itself and data harvesting companies in several countries (I suppose most) are even obligated to hand out information if it would assist in an investigation.
What's this gotta do with a compression format!?
Coming from one of those major companies the compression format in question is written in C and published under the BSD License and GPLv2 on Github. Connecting this with the former paragraph it is rather apparent that this might empower this specific company to become more and more a major part of the Linux community. Linux has grown and matured a lot and even gained user base on personal computers and it appears companies are getting aware of it. Also see Fungalnet's article on his blog about this where he mentions the recent purchases of Github and RedHat.
Being a part of something also means a certain influence and possibilities to exploit that something. This is true for our very life and any social construct to begin with. In this situation it may open way for Facebook to inject questionable code to harvest data into Linux. This is basically what this is all about.
My very personal opinion
Disclaimer: This is my personal opinion based on the little information I have on this matter. Please correct me where correction is due. Also I'm too lazy to search for sources here, as most of this is just very basic stuff.
To me personally the reason behind putting money and effort into the development of a multi threaded compression algorithm would suggest interest in pushing faster mobile and embedded device operation and gaining market share by being an active force in said development (see concerns and what that's gotta do with anything). AMD and Intel are fighting heavily for market share on PC/professional with immensely powerful work station and gaming CPUs getting smaller and smaller while ARM is already dominating the mobile market with just as powerful and more energy efficient processors for micro environments, which in reality is way bigger nowadays anyways.
For me there's several sides to this:
- The zstd code is completely open source and could be checked for any questionable hooks or back doors at any moment. Although people might do that at first and later on get lazy and not get aware of such occurrence until it had already done some damage.
- In a time where wireless radio frequencies are so fast that you can download a 10GB file in a few minutes (not where I live anyways) and processors that render almost life realistic video scenes in a matter of a few hours it seems absurd to opt for a compression gain in a package manager for desktop systems.
- That being said Arch is not only desktop. It's also server, embedded and maybe even mobile in the future.
- It is without a doubt a disruption for many Arch based distributions and Arch forks and implies decisions and work for their developers and maintainers
- I've mentioned this in the causing thread but I'll repeat it here: Although I love to use, collaborate and share open source efforts and the general idea of working together to create great things I'm also aware that I'm looking at a (pretty old) LG monitor, listening to music from Spotify on (pretty old...) JBL stereo speakers using an Intel processor on an MSI motherboard with a SAMSUNG solid state drive. Hell even the enclosure all this resides in is not open source. It is made by Thermaltake. All these have in common that they make business, have business interests and make business decisions which in return empowers them to invest in doing more. I could not run Arch if it wasn't without any of this. Here's something to think about: There have been and still are open source solutions to processors, components, full systems like laptops and others yet most of them find themselves in a difficult situation. Plain business oriented companies can invest loads of money to advance and develop their stuff which is what everybody wants! If we stopped wanting more power, more screen real estate, more sound quality (I don't even see what for...), more speed etc. then we could start looking at alternatives. It is one major drawback of Linux/UNIX and all their derivatives that the relatively small developer base and private effort of a few without monetary compensation can not create the same in the same amount of time. (Although it can sometimes create even greater things in less amount of time). When it comes to sheer numbers, which our society claims for, quantity just wins over quality. No news here either. And if you're honest to yourself a lot of Linux development (as in evolution) happens because its major market is server and embedded and there is business decisions in this fact alone because companies want to use it for their benefit not for their spare time.
Discuss, share your views, your knowledge, your opinion!
( In a civilized manner apparently ;) )