FSF-Approved Hyperbola GNU/Linux Forking OpenBSD, Citing 'User Freedom' Concerns (hyperbola.info) 135
Long-time Slashdot reader twocows writes: Hyperbola GNU/Linux, a FSF-approved distribution of GNU/Linux, has declared their intent to fork OpenBSD and become HyperbolaBSD..."
The news came earlier this week in a roadmap announcement promising "a completely new OS derived from several BSD implementations" (though Hyperbola was originally based on Arch snapshots and Debian development).
"This was not an easy decision to make, but we wish to use our time and resources to create a viable alternative to the current operating system trends which are actively seeking to undermine user choice and freedom." In 2017 Hyperbola dropped its support for systemd -- but its concerns go far beyond that: This will not be a "distro", but a hard fork of the OpenBSD kernel and userspace including new code written under GPLv3 and LGPLv3 to replace GPL-incompatible parts and non-free ones.
Reasons for this include:
- Linux kernel forcing adaption of DRM, including HDCP.
- Linux kernel proposed usage of Rust (which contains freedom flaws and a centralized code repository that is more prone to cyber attack and generally requires internet access to use.)
- Linux kernel being written without security and in mind. (KSPP is basically a dead project and Grsec is no longer free software)
- Many GNU userspace and core utils are all forcing adaption of features without build time options to disable them. E.g. (PulseAudio / SystemD / Rust / Java as forced dependencies....)
HyperbolaBSD is intended to be modular and minimalist so other projects will be able to re-use the code under free license.
The news came earlier this week in a roadmap announcement promising "a completely new OS derived from several BSD implementations" (though Hyperbola was originally based on Arch snapshots and Debian development).
"This was not an easy decision to make, but we wish to use our time and resources to create a viable alternative to the current operating system trends which are actively seeking to undermine user choice and freedom." In 2017 Hyperbola dropped its support for systemd -- but its concerns go far beyond that: This will not be a "distro", but a hard fork of the OpenBSD kernel and userspace including new code written under GPLv3 and LGPLv3 to replace GPL-incompatible parts and non-free ones.
Reasons for this include:
- Linux kernel forcing adaption of DRM, including HDCP.
- Linux kernel proposed usage of Rust (which contains freedom flaws and a centralized code repository that is more prone to cyber attack and generally requires internet access to use.)
- Linux kernel being written without security and in mind. (KSPP is basically a dead project and Grsec is no longer free software)
- Many GNU userspace and core utils are all forcing adaption of features without build time options to disable them. E.g. (PulseAudio / SystemD / Rust / Java as forced dependencies....)
HyperbolaBSD is intended to be modular and minimalist so other projects will be able to re-use the code under free license.
Lotta hype ... (Score:3)
Sorry.
Low-hanging fruit.
Re: (Score:2)
Re: (Score:2)
If they are going to use a code of conduct, they might have to!
Look what they did to Linus...
Somehow I still think Theo would tell them to fsck off:)
Re: (Score:2)
Still, I like the idea. I always felt like Linux was becoming too much Windows like with systemd and something else based on BSD would be refreshing since we are once again moving towards a monopoly lately, albeit open sourced. Next, let's fork that ceremier guy achievements, the situation sure couldn't get any worse than what it is now so that would definitely be a win.
Stop trying to fork me over.
Re: (Score:2)
Well, based on UNIX. Starting from BSD licensed code.
To me it is quite interesting, though that's not the area of coding I work in. If UNIX could properly handle ext3 file systems I would have tried it out already. As it is, the last time I used UNIX was on an i386 based system licensed from AT&T, so I've no real idea what modern UNIX systems are like.
Re: (Score:2)
FreeBSD supports zfs out of the box, which is way better than ext3 or anything else that Linux supports, including its own implementation of zfs.
Re: (Score:2)
"You don't need your data, here, have some other data. It is better, I promise!"
Re: (Score:2)
Last time I checked that was a read-only driver. But, yeah, I should have said ext4. It's been awhile since I checked my disks, and I was remembering an old system.
Thing is, I don't have a spare computer, so I'd need to keep my main partition accessible from both systems. (They'd need to share the disk.) And that's a bit nerve wracking. And most of the time I'm not mad enough at the current system to want to just junk it without knowing I've got a replacement that's viable for my purposes.
I'll get the popcorn (Score:2)
Re: I'll get the popcorn (Score:2)
Re: (Score:2)
In Unix and Linux it's spelled kernel.
HDCP Support (Score:4, Interesting)
Do a quick Google for "van Eck phreaking," and you'll see that your desktop sessions can be snooped wirelessly via the EM radiation coming off your VGA/DVI/HDMI/DisplayPort cable, or by EM radiation from the monitor itself (albeit only with rather sophisticated equipment in close proximity). It occurs to me that HDCP may help thwart this.
So: Make HDCP an administrative security setting. Unprivileged applications don't get to change it, and have only limited ability to inspect it. If an unprivileged application wants it turned on, a security dialog pops up, and the user gets to decide whether or not to allow it. In other words, HDCP is re-contextualized to serve the security interests of the user, and not the entitled twerp who thinks they can run roughshod over your computer in exchange for a few minutes of mass-produced entertainment.
Just a thought...
Re:HDCP Support (Score:4, Informative)
"Do a quick Google for "van Eck phreaking," and you'll see that your desktop sessions can be snooped wirelessly via the EM radiation coming off your VGA/DVI/HDMI/DisplayPort cable, or by EM radiation from the monitor itself"
In short, that is enormously difficult, and your monitor cable is shielded, so realistically it has to be done from your display, which has already decoded HDCP. In order to realistically mount an attack of this type, you have to be very near the target, preferably just the other side of the wall. You'd be better served by hanging up some aluminum foil and grounding it than depending on HDCP.
tl;dr (Score:2)
You're not important enough for Uncle Sam to bother with.
Re: (Score:2)
Indeed. And van Eck phreaking is actually pretty expensive, tedious and has uncertain and variable results.
Re: (Score:2)
That hardly matters when the attacker can just make things up. Spies aren't even needed anymore.
Not difficult! (Score:2)
A PC with a TV card in the room next to it
you is enough.
Which, in an apartment building, or hotel, etc is easily the next apartment over. Or in front of the window, in a car. Or in the hallway.
And the whole point is that the built-in shielding is not enough.
While wrapping everything in "tin" foil only achieves making you stand out as suspicious and look lime an idiot.
Also, the old "Why would they spy on ME" nonsense... Why would they exclude you and NOT spy on you, specifically?? It is easier to just spy on
Re:HDCP Support (Score:4, Informative)
AFAIK HDCP keys are hardcoded, so once an attacker has the HDCP keys, which have leaked of course, the security advantage goes away. With that in mind, I don't think you can call HDCP a security setting.
Firm-coded, at best! (Score:3)
They don't etch a different cirquit for each one, you know?
It's on a built in Flash-like memory that holds the firmware. It changes with updates. Hell, for some hardware, it gets loaded from hard disk on every boot, because they saved on the "EEPROM" chip.
It is trivial to code a key setting utility.
Re: (Score:2)
You'd need to sniff the initial negotiation to get the stream keys using the known hard-coded root keys. You'd need to catch the moment the monitor is switched on, or comes out of sleep. Anyway, as I said in my other comment, it's far more practical to capture the radiation from the panel itself rather than from the cable.
It's the LCD panel itself (Score:5, Informative)
The grid of electrodes in the LCD panel itself acts as a big radiating antenna. The video has obviously been decrypted, unscrambled, and decoded to RGB by the time it gets there. The radiation from the cable is tiny by comparison.
Wasn't BSD always more free than Linux? (Score:1)
Re: (Score:2)
Depends on whether you use a libertarian's or a socialist's definition of "freedom." For the former, yes, for the latter, no.
Re: (Score:2)
When debates about definitions get emotional, they lose sight of the fact that definitions are mere *conventions*. Group A says "Freedom is X and Freedom is Good"; Group B says "Freedom is Y and Freedom is Good," and we automatically act as if one and exactly one of them must be right.
Logically it'd be the same if A were saying, "Fizbo is X and Fizbo is Good," and B saying "Luzbat is Y and Luzbat is Good." Neither of them is necessarily right, nor if one of them is right is the other necessarily wrong.
Re: (Score:2)
When debates about definitions get emotional, they lose sight of the fact that definitions are mere *conventions*.
Possibly, but we also have the case where some definitions are simply erroneous. As we have here where on group turns a blind eye toward the *restrictions* that *enforce* a desired benevolent behavior.
Re: (Score:2)
A definition can't be "erroneous". It can be *non-standard* (not mean what most people think it means). It can useless (e.g. refer to something that can't exist) or absurd (assume the existence of things that aren't compatible with each other). But the one thing a definition can't be is wrong.
When you argue about definitions, you are arguing about terminology, not reality.
Re: (Score:2)
I find the big problem with definitions is that they don't correspond to what people think they do. They think "Freedom of Speech" means they can say anything they want without consequence. But they don't think other people should be able to. This isn't a problem with definitions, it's a problem with thinking. You could define "Freedom of Speech" that way, but very few people would want it. Mostly they use terms like that as a glittering generality.
Re: (Score:2)
Language is absolutely "some amorphous blob," and if the language is English that is explicit.
All the authoritays of the English-speaking world got together for big conventions, multiple times in history, with the goal of standardizing the language. The result each time was that everybody changed sides from "yes we want to standardize" to "no, never" after they got together and found out what each other wanted.
Completely different than other European languages, which usually succeeded in their standardizati
Re: (Score:2)
Definitions exist, but they don't come from an authority. They're descriptions of the known uses, according to somebody, not lists of approved meanings.
It is not incorrect to say that most words are cromulent. It was not incorrect at any time. When Lisa Simpson offered the definition, nobody actually even knew at that point if she was defining a real word, (she was) or if nobody else would use her definition and it would not be a real word.
The appearance of agreement when people are communicating is much st
Re: (Score:2)
Congratulations, you nearly comprehended a simple point, but you were so unsure if you had it right you sarcastically questioned your understanding.
Consider next the utility of using different definitions than everybody else, and if it has any significant impact on the others, or mostly just on you.
Not a matter of convention. (Score:3)
The libertartian "freedom" to take other people's freedom is never accepted. Not even by libertarians. It is essentially just a disguised newspeak word for them wanting to be dicks and harm others for their selfish reasons, while being too damn stupid to realize there is always somebody who is going to do that to *them* too. :)
Re: (Score:3)
You do know that libertarian was originally a socialist ideal?
From https://en.wikipedia.org/wiki/... [wikipedia.org]
Re: (Score:3)
Either way, it's not a particularly relevant or useful argument. Modern libertarianism and modern socialism are generally well-understood and distinct schools of political thought. Even if the origins of modern libertarianism did trace back to socialism so
BSD is more *permissive*, not more free (Score:3, Insightful)
Formally, the BSD license is a more permissive [wikipedia.org] software license than any version of the GPL, but not necessarily a more free license since BSD licensing may lead to loss of freedoms. This difference between the licenses extends to BSD vs Linux distros as well. The linked article provides full details of why degree of permissiveness is not the same thing as degree of freedom, primarily in its impact on developers versus users.
The same situation is seen in other social settings. For example, imagine a
Re: (Score:2)
The problem with your example is that one freedom is infringing on someone else’s freedom, or to put it in libertarian words, your freedom ends where my property begins. There are also people that think a base morality supersedes the freedoms available.
Re: (Score:2)
I've personally never forgiven Stallman for adding "compelled public benevolence" to the dictionary under "freedom".
If I make a USB flashlight using wire-wrapped TTL, it has no source code, and I haven't deprived anyone else of their "freedom" (whew).
If I make the same USB flashlight with a PCI10 and a several dozen instruction
Re: "Failure to gain" is not a "loss" (Score:2)
Re: (Score:2)
Your analysis is incorrect.
When a previously-open package is turned into a closed proprietary one (as permissive licenses allow), the great majority of its users suffer a loss of freedom, because the pre-closure version of the package becomes unmaintained.
Again, you confuse a failure to gain something new, in this case bug fixes and new features, with a loss. They lost *nothing*. The formerly open software they were using is unchanged. Its works as well as it did before the proprietary version.
In today's security-hostile environment, that turns the old version into an immediate liability ...
Again, the liability is not something new. They was always a security problem. You confuse being blissfully ignorant of the liability with not having the liability. Nothing was lost other than ignorance of a problem, which is something other than freedom.
In fact, the users becomes hostage to its developer through the need for security fixes ...
No they do not
Re: (Score:2)
There's an interesting little logic problem entailed here.
Is the way to maximize the freedom enjoyed in the world to allow the most freedom to the people you deal directly with? Many people would say yes, but some would say we need to restrict the ability of people to restrict other peoples' freedom.
It's an interesting philosophical debate, but in practical terms both approaches seem to work.
Linux/GPL "benevolent" not "free as in freedom" (Score:2)
Wasn't BSD always more free than Linux?
Yes. Linux, well the GPL actually, has restrictions. The fact that those restrictions are designed to be benevolent does *not* change the fact that they are restrictions. Linux, like BSD, is free as in beer, but Linux is less free as in freedom.
Please, before typing in your 1000 word manifesto, re-read the above, and then contemplate whether you are conflating benevolent behavior with freedom. Hint: they are two separate things. The restrictions that enforce benevolence are inherently less free as in fre
Re: (Score:2)
The question isn't benevolence, but scope. For the individual with BSD licensed code, their actions are freer. But for the recipient of that code...being able to re-license it, or to deny access to the source code, decreases the freedom. Motives are beside the point in this argument.
Re: (Score:2)
The BSD license allows anyone to do almost anything with the software. (You aren't allowed strip off or change the notes on who wrote the software; and some versions of BSD have an "advertising clause" that requires you to disclose where you got the software if you use it. Other than those points you can do what you want.)
The GPL license family puts some restrictions on the software, and GPL proponents believe that these restrictions maximize freedom for the community using the software.
For example, with
Linux code is more free than BSD code (Score:3)
You should make the distinction between freedom of the code itself and freedom to use it. GPL relates to the freedom of the code. That is the intent is to get sure the code remains free. Of course you may freely use the code however this comes with the associated burden to give back any improvement you've made. Whereas the BSD license applies to the freedom to use the code. And why not change it and then render it non free with a non free license. Which means the original BSD code won't benefit from said ch
Re: (Score:2)
It is, it's just not communist like the gpl
Re: (Score:2)
Depends on how you look at things. BSD licensed code is more free to the person who takes it, but may well be non-free to the person who receives it from them. GPL imposes transitive-rules on freedom, but BSD doesn't. So if you are dealing with lots of code, much of which is binary, then code with BSD licensed sources is likely to be less free to you. But if you're working with only source code, BSD licensed code is freer, in that it doesn't require that freedom be transitive.
Re: (Score:2)
Yes, but that freedom comes at a cost, which is why these people are going to be replacing as much as they can with GPL'd code.
I'd be curious to see an attorney who's familiar with FOSS licenses comment on it. My personal take on this is that nobody will use a BSD kernel that's been infected with GPL code.
Public domain (or equivalent licensing) is great for achieving certain goals, but there's a reason the Linux ecosystem has been much more vibrant.
I'm pretty sure it's not the license. IMO the Linux community is better organized – period. That's the sole reason Linux is as big as it is. If Bill and Linn Jolitz had been more welcoming the BSDs would never have forked and BSD might have been every bit as big as Linux is.
Re: (Score:2)
My personal take on this is that nobody will use a BSD kernel that's been infected with GPL code.
Linux kernel users will, in particular those that fear the Linux kernel is hitting that software lifecycle wall where maintenance is getting increasing difficult, and/or the bloat is getting beyond the minimalists tolerance level, and/or for those that believe philosophical purity is the one true choice.
They already use Linux. Your argument is moot. (Score:2)
So I don't think they would call a license that doesn't allow leeches to leech off of you and TiVoize your projects an "infection".
Excluding such leeches is kinda the point of the GPL.
This isn't a project for leeches and nutjobs like you. They can fuck right off.
Re: (Score:3)
... but there's a reason the Linux ecosystem has been much more vibrant ...
True, its called the "network effect" and it has little to nothing to do with the GPL. In the early 1990s Linux distributions crashed less than FreeBSD distributions. So people used and developed for Linux. People wanted a viable *nix for commodity PC hardware, few cared about the politics. This gave Linux the early "network effect" win.
Re: (Score:2)
It had less to do with crashes and licenses than with BSD being tangled up in legal issues preventing its distribution. Linux got a several year running start while BSD was getting those straightened out. By that time Linux had become the default.
BSD is hardly doing badly, considering it's included with every Mac. When it comes to desktop unix, BSD installations massivley outstrip Linux, courtesy of Apple.
Re:Wasn't BSD always more free than Linux? (Score:4, Insightful)
It had less to do with crashes and licenses than with BSD being tangled up in legal issues preventing its distribution. Linux got a several year running start while BSD was getting those straightened out. By that time Linux had become the default.
My perspective was 1993'ish. FreeBSD was out, just less reliable.
Reliability, legal issues, yeah. GPL, no, that's fanboy fantasy. People just wanted Unix on their PCs and BD vs GPL was irrelevant to nearly all.
Re: (Score:2)
Hyperbola or Hyperbole? (Score:2)
Re: (Score:2)
hopeless (Score:2)
Small team is trying to hard fork a whole operating system and improve it. It will end up as yet another GNU Hurd - toy project without any practical use. Whats even worse they intend to GPL3 their code, so even the potentially useful code they create will remain useless for general public as it wont be possible to port it back to *BSD due to license conflict.
Re: hopeless (Score:2)
Re: (Score:3)
Which is why some people really hate the GPL, because GPL fanatics always complain "but someone can steal your code!", yet completely ignoring the fac
Comment removed (Score:4, Interesting)
Re: (Score:2)
Re: (Score:2)
You don't get forking, do you? (Score:2)
The smallest fork is ... a pure name change.
Why would forking mean any more work than you want?
The base system won't care.
I can run a fork all on my own, in my spare time. The weekly routine (mostly automated) goes like this:
1. Update my fork with all the changes from the original.
2. Apply my patches.
3. For those that fail, and specifically the lines that fail, do the changes manually and then do a diff to get a new complete patch.
4. Maybe code a bit on my own parts, to add a feature or fix a bug.
5. Commit/
Re: (Score:2)
They intend to improve security, remove "tainted" dependencies ("E.g. (PulseAudio / SystemD / Rust / Java as forced dependencies)"). This definitely doesn't look like a small fork with minimal patching but more like a serious rewrite. They even clearly state "we are planning on implementing a completely new OS derived from several BSD implementations.".
Re: (Score:2)
They want to use GPL3 licensed code. This means that they can't use Linux modules that don't have that "any future version" line in the license.
If you want to know what's allowed, anything that's allowed by the GPL3 license will be allowed. And possibly a few things that aren't, but which only use subsystems, but don't depend on that.
This is no problem for me, as I prefer the AGPL3 license, and the difference from GPL3 rights isn't something that affects my work. But if you're planning to take it, repack
That's one poorly-written submission title (Score:3)
"FSF-Approved Hyperbola GNU/Linux Forking OpenBSD, Citing 'User Freedom' Concerns"
This certainly sounds as if the concern is with OpenBSD - however, reading the summary, it's obvious this is being done because of concerns with the current state of Linux's kernel, and they see the OpenBSD kernel as a better option.
Taking a pragmatic point of view - I'd think their time would be better spent leaving the OpenBSD kernel alone and focussing more on addressing (what some might perceive as) the lack of a quality UI for OpenBSD, which might make it more appealing to a broader audience. I certainly have more faith in the OpenBSD kernel maintainers than I do in these people's ability to rewrite that kernel securely. But it is their own time to spend as they see fit, and I realize pragmatism doesn't weigh heavily in decisions like this.
Re: (Score:2)
And we're "concerned" about this? Really? Have you checked the calendar recently? Like in the last two decades or so? Last *I* checked, even some of the twats from systemd weren't allowed Linux kernel checkins (any more).
Re: (Score:2)
a centralized code repository that is more prone to cyber attack and generally requires internet access to use.
My take was that they were talking about Rust, and the possibility of it being used in the kernel. I only read the summary.
Re: (Score:2)
We arent talking about userland here. We are talking about the kernel. The kernel, for lack of a better term, is sacred. It is not a place to smear shit.
Re: (Score:3)
Even the act of floating the idea that Rust should be used for kernel code should get your kernel contribution rights immediately and irrevocably revoked.
OK, now we know that you are of that opinion. But where's your argument? If you want to make the case that Rust does not belong in the kernel, you should present some reason as to why. "Everyone who disagrees with me should be kicked out" is not a compelling argument.
Re: (Score:2)
Re: (Score:2)
2. What's the fix or alternative, and why is that better?
Re: (Score:2)
1. It's nearly the worst successful language to succeed assembler because of it's tendency to memory leaks and the difficulty of restricting wild pointers.
2. Ah...the alternative is...
Damn. Everything else is clumsier or larger or less efficient.
Re: (Score:2)
Re: (Score:2)
Well I wish them success, it looks like a lot of work though. But I can understand their concern, seems Linux is being dragged/changed into a corporate controlled project.
I hope they keep the OpenBSD's rabid obsession with up to date man pages and launch anything even remotely associated with info(1) into the Sun or better yet a black hole
Re:That's one poorly-written submission title (Score:4, Insightful)
Sounds like they are not UI oriented. And you are asking them to be UI oriented. While you admit it is their time to use, it may be a team who should in no way be making UI decisions because of their interest and talent being directed elsewhere.
Re: (Score:2)
Thought I was having a stroke while trying to parse it.
Re:That's one poorly-written submission title (Score:4, Insightful)
Who says OpenBSD lacks a quality UI? Doesn't it run KDE?
The reason most of us choose Linux over BSD is that most people choose Linux over BSD. Being more popular means better hardware and software support. (Same reason a lot of people like Windows.)
Re: (Score:2, Funny)
Who says OpenBSD lacks a quality UI? Doesn't it run KDE?
Yes it does run KDE. Yes OpenBSD lacks a quality UI.
Re: That's one poorly-written submission title (Score:3)
Hardware support is the rub (Score:4, Informative)
Part of Linux's success is its support for practically all hardware. The only reason I don't use OpenBSD is that it doesn't support all the hardware on the machines I want to run it on, although I haven't tried in a year or two. Someone actually contributed a patch to the relevant atheros driver to support my particular nic but they rejected it on the basis that it might violate the GPL... but the patch only used values from the Linux driver, no code, so it has already long been settled that it would not violate the GPL. Enough time had passed that the patch would no longer apply cleanly, and enough had changed that it was not trivial to patch manually... So I installed Linux.
OpenBSD is not just an OS which is not for everyone, it is WILLFULLY not for everyone. Which is stupid.
If this fork somehow has better hardware support, I'll use it. But that seems unlikely...
Re: (Score:2)
In this case, the makers insist that certain hardware and features should not be supported (eg. HDCP) because they do not agree upon freedom.
For me, it would depend on the project, if I’m building something that has to be able to be maintained between now and hundreds of years after my death, I would want something like this project that has guaranteed Intel or Microsoft will never make claim to my creation.
If I want something that just works, I’ll go with a Debian distro without any of Poetteri
Re: Hardware support is the rub (Score:4, Informative)
It is stupid, to not turn it into another iOS? (Score:3)
Alright then! Keep being used by, err, using your iPad.
We'll stay with the pro tool, for people who know what they are doing, that will take your foot off if you tell them to.
And you don't get to ruint it for us, by turning it onto another piece of condescending simplified cumbersome shiny crippleware.
But you're too stupid to get to go around calling other people stupid. The fact that you can't even write drivers should already be enough to tell you that.
Go back to your iPad.
Re: (Score:2)
It's the pro tool because the developers willfully decided not to support another flavor of the same wifi NIC that they already supported? Okay, sport.
Just use NetBSD with Pkgsrc (Score:3, Informative)
Why roll something new that will need to be maintained as a 'distro'? You can just install NetBSD and use the fairly large pkgsrc collection for user applications. The linux 'distro' brings baggage into things that is unnecessary.
Re: (Score:3)
Why not use NetBSD? Because even its own key developers abandoned it in the 1990's, shortly after Theo de Raadt was kicked off the project in roughly 1995 and founded OpenBSD instead. I admit that I will be very surprised if OpenBSD's current developers will accept code or cooperate with this FSF licensed and GPL licensed work.
Re: (Score:3)
NetBSD has a troublesome maintainership, in the sense that while its in theory portable to almost everywhere, many of the target platforms and ports are now subtly broken, with insufficient maintainers to fix it. It does have very modern ports in pkgsrc, but Xfce won't correctly run on it currently, and that is an achievement. I actually came to love fluxbox because of this, though. If NetBSD got more love, yes, it would be great again, but right now it is below critical mass required to be maintainable.
On security... (Score:1)
- Linux kernel being written without security and in mind. (KSPP is basically a dead project and Grsec is no longer free software)
This is not a linux problem; it is the result of legacy architectures where security was a belated afterthought, motivated by 32-bit limitations. The multiple address space hack with page granularity protection is manifestly insufficient, making IPC excruciatingly slow and painful. Until hardware is more capable, security and performance will continue to be mutually exclusive. Where performance matters at all, kernels and applications will remain monolithic and untrustworthy.
Re: (Score:3)
Security and performance will ALWAYS be mutually exclusive - look at Rust or Java for examples - because both are relative and we’ll always need/want the bottom of the can as for performance.
What can help is code-time performance and security analysis. There’s plenty of static and dynamic ways of testing code but writing unit tests is simply not the way to go with that and although some automated testing is possible, the really good stuff is still locked behind massively expensive commercial gat
Obligatory "NetBSD is Dying" post (Score:3)
One more crippling bombshell hit the already beleaguered NetBSD community when IDC confirmed that NetBSD market share has dropped yet again, now down to less than a fraction of 1 percent of all servers. Coming close on the heels of a recent Netcraft survey which plainly states that NetBSD has lost more market share, this news serves to reinforce what we've known all along. NetBSD is collapsing in complete disarray, as fittingly exemplified by failing dead last in the recent Sys Admin comprehensive networking test.
You don't need to be a Kreskin to predict NetBSD's future. The hand writing is on the wall: NetBSD faces a bleak future. In fact there won't be any future at all for NetBSD because NetBSD is dying. Things are looking very bad for NetBSD. As many of us are already aware, NetBSD continues to lose market share. Red ink flows like a river of blood.
FreeBSD is the most endangered of them all, having lost 93% of its core developers. The sudden and unpleasant departures of long time FreeBSD developers Jordan Hubbard and Mike Smith only serve to underscore the point more clearly. There can no longer be any doubt: FreeBSD is dying.
Let's keep to the facts and look at the numbers.
OpenBSD leader Theo states that there are 7000 users of OpenBSD. How many users of NetBSD are there? Let's see. The number of OpenBSD versus NetBSD posts on Usenet is roughly in ratio of 5 to 1. Therefore there are about 7000/5 = 1400 NetBSD users. BSD/OS posts on Usenet are about half of the volume of NetBSD posts. Therefore there are about 700 users of BSD/OS. A recent article put FreeBSD at about 80 percent of the NetBSD market. Therefore there are (7000+1400+700)*4 = 36400 FreeBSD users. This is consistent with the number of FreeBSD Usenet posts.
Due to the troubles of Walnut Creek, abysmal sales and so on, FreeBSD went out of business and was taken over by BSDI who sell another troubled OS. Now BSDI is also dead, its corpse turned over to yet another charnel house.
All major surveys show that NetBSD has steadily declined in market share. NetBSD is very sick and its long term survival prospects are very dim. If NetBSD is to survive at all it will be among OS dilettante dabblers. NetBSD continues to decay. Nothing short of a cockeyed miracle could save NetBSD from its fate at this point in time. For all practical purposes, NetBSD is dead.
Fact: NetBSD is dying
GNU Hurd is dead! All Hail GNU Hyperbola! (Score:3)
Purism is the fastest way known to irrelevance
Forks are Good. (Score:2)
Re: (Score:2)
the HURD (Score:2)
whatever happened to that project ?
Re: (Score:2)
Too Hurd to get it to wurk!
Re: (Score:2)
Dunno, haven't hurd anything lately
April's fools? (Score:2)
I didn't notice we were already nearby 1st of April.
Also: Linux is failing at getting the ideas of Uni (Score:2, Interesting)
With the worst offense being that not everything being a file anymore, merely because they were too dumb and lazy for it. ("...don't know how to do it..." and "...difficult...", according to Torvalds)
Making usage of those parts cumbersome, inelegant and unnecessarily complicated.
Strange then, that others have no problems giving them a file system interface. And you could have copied them.
Also, more and more tools seem to fail at basic fearures like providing simple doxumented text configuration files, scrip
Re: (Score:2)
I think it's more the Bazaar vs the Cathedral combined with Theo's uncompromising stance on quality proving to be the better model.
BSD isn't Linux, BSD is an OS. Linux is a kernel.
I wonder how long it takes folks to figure this out ? ( been a few decades already )
Re: FSF is dead (Score:2)
Re: (Score:2)
I echo your sentiment on finding talent.
I think we need to push people to learning how to learn instead of googling for howtos.