Hacker News Reader

Comment by traceroute66 | Hacker News Reader

OpenBSD is great. Except for Theo.

If you question Theo on anything, even something reasonable, you will incur his wrath, and that of his henchmen (the clique of project devs that surround him).

Case in point is the whole "Funding our Electricity" thing: (a) Theo insists on keeping their test lab in his own house (b) Theo insists on supporting obscure archaic architectures (c) Theo claims on list to have incurred 20,000 dollars of electricity that year (d) Theo refuses to answer reasonable questions, such as cost breakdown, why the equipment can't be moved elsewhere, why the insistent on obscure architectures. Or when he does answer its essentially "my way or the highway".

OpenBSD is great, no doubt about that. But the whole Theo & friends clique needs to change.

View on HN

"my way or the highway" is the best part of open source organization.

It's not democracy. There is no equality. You have no voice unless others want your contribution.

People are free to form a cliques of same minded people and ignore everyone else. You can hate those who make the software you use, but Theo & friends do what they want.

I think that's the great thing about voluntary association and freedom in that sense in general: you can associate with those who are like-minded and willing to work with you, and you don't have to/aren't forced to associate with those that disagree with you entirely - and those groups can't really get at you as easily. In open source, it's just a more pure version of that because you essentially have infinite "space" to spread out, and you can just clone stuff.

Not directly relevant to this discussion, but I think it is worth pondering about:

While I greatly enjoy the freedom that comes with voluntary association, I think it is worth realizing that it has few protections against the creation of toxic environments that exclude people/resources on non-technical issues. Our culture can be annoyingly straight-white-male oriented.

I don't think you can easily separate a great project like OpenBSD from the culture that produced it. OpenBSD is what it is precisely because Theo and co. are ruthlessly uncompromising with their vision and their principles, and don't suffer fools. That may for some unpleasant email exchanges and hurt feelings, but OpenBSD is a great and unique OS and has been around for decades.

Perhaps. But it's also difficult to separate one's donations from a desire that those funds are going places where they will be handled appropriately.

The OpenBSD Foundation tells you how your funds are going to be used, if you instead choose to donate to the project, then you're donating to Theo and it is entirely his discretion how he chooses to spend it on the project. Donations with strings attached are not donations.

https://www.openbsdfoundation.org/activities.html

vs.

https://www.openbsd.org/donations.html

I don't really understand what the point (b) is about. Even if we assume, that portability in and of itself isn't a good thing (which, I think, it is), as far as I know, OpenBSD is very popular in routers and other network-related devices, where “obscure archaic” architectures are not as obscure and archaic.

I think this is his equipment rack in his basement from 2009. The image is from the official OpenBSD web site.

http://www.openbsd.org/images/rack2009.jpg

Security is an important issue to them so they feel they need physical direct control of the hardware and don't want it located somewhere else.

Also some of the equipment is old and may not survive being transported.

I remember them saying taht alignment and endian issues can only be found when compiling and running on the actual hardware. Cross compile to a different arch and emulators don't catch the same bugs.

No wonder someone mentioned how expensive his electricity bills were. Some of those older machines eat some power. I was looking at old Silicon Graphics servers but the noise and power consumption was too much.

That's some "cops kick down the door to shut down your illegal marijuana grow operation" level of electricity use.

Looks like a fire waiting to happen. Electricity, heat, small space, wooden beams, carton boxes.

I have read various times about this thing of Theo being angry at people and so on which I don't understand honestly. I know his replies are usually short and he doesn't like to hold hands to n00bs like us, but I never had the feeling he was angry or insulting people, it was quite the opposite: his mail are on spot, precise and, frankly, honest.

Disclaimer: I've been using OpenBSD since 6.3 (~ 2 years ago), so quite recently, I don't know if he changed attitude at some point.

I lurk the OpenBSD mailing lists and while Theo can definitely be abrasive, I think that it's really that he won't suffer fools lightly.

What seems to be the least tolerated are "ideas people". If you come in and suggest that it'd be lovely to port OpenBSD to this new SOC, or that someone should write new Bluetooth support, or what have you – that's when the teeth come out.

But conversely, if you bring a patch the OpenBSD community seems to be very welcoming. They definitely value you putting in the work, not just talking about it.

For those critical of his personality, imagine being constantly trolled by script kiddies and others with malicious (such as government actors) intent. You'd probably get a little grumpy, too. The fact that Theo and the OpenBSD crew have been so uncompromising ideologically is a large part of its attraction.

The support for "antiquated" architectures has a lot to do with them being OPEN, e.g. SPARC. This crew is some of the first along with rms to really push not just open and free software but also the hardware it runs on for a variety of solid reasons.

Well...OpenBSD runs on a lot of "antiquated" architectures because it started as a fork of NetBSD, which supports/supported even more architectures. Both Open and Net have pared down the definition of 'supported' and the number of archs they support over the years.

The support for "antiquated" architectures has a lot to do with them being OPEN

Not exactly...I recall Theo called it OPEN because he felt the NetBSD crew wasn't. NetBSD was around first, recall, and one of their core principles is wide portability, where OpenBSD took on the mantle of correctness and security and portability came along for the ride. Both side of the unpleasant history are documented; the truth is somewhere in the middle, of course.

That said, both crews advocate building on multiple architectures as a solid best practice to turn up 'interesting' programming issues. For example, Theo has stated that SPARC/SPARC64 porting is particularly useful for finding unaligned memory accesses (Alpha didn't like them either).

I'll leave out my own experience in dealing with the 'OpenBSD attitude' other than to say 'generally less than pleasant'. I still use OpenBSD daily, tho more for infrastructure (FW, DNS, DHCP, LDAP) than as a desktop.

From what I've seen on the mailing lists, he seems a lot like Linus, where he's only rude if he thinks you should know better, or if you make an objectively stupid suggestion. I've regularly seen contributors being rude and insulting though. For example, I read a message from one contributor who said that safety features in modern languages are like a straitjacket, and are only useful to developers too stupid to write bug-free programs in C.

>OpenBSD is great, no doubt about that. But the whole Theo & friends clique needs to change.

Why -- he effectively builds and designs it and it's open source and free to copy and make your own.

Those architectures help OpenBSD to smoke out bugs.

Obscure architectures is important also to keep the world online.

Also obscure architectures are damn fun.

I bet the disappearance of big endian systems means a notable amount of bugs sit out there waiting.

Are there advantages to big endian systems?

Network protocols usually encode things as big endian since that was the dominant order when the internet was being developed. It's a minor point now but everything has to be swapped around on little endian systems.

But is there some kind of computational advantage down to the processor architecture level? Can certain instructions or set of instructions be executed faster on big endian systems?

No. There may have been micro optimizations on specific chips, but there's nothing intrinsically faster about one encoding or the other.

Sorta like English reads right to left, but left to right or top to bottom would work just as well. left to right was what won.

No, no advantage. Just variety. You used to occasionally see bugs where things would work fine on little and not big, or vice-versa. Usually due to assumptions about data.

Not sure about today's CPUs, but back in the days of the 68000 I'm pretty certain there was a small advantage in both designs but for different reasons. I wouldn't normally write such a vapid post without supporting details (that I can't recall anyway), but I wanted to pipe up to encourage you to do more research if you're curious.

I would imagine the new PowerPC64 port is running big endian.

Cool! Do you know which machines that port would boot?

http://www.openbsd.org/powerpc64.html

Current status:

Currently OpenBSD/powerpc64 only supports POWER9 processors and runs stably on PowerNV machines based on the Raptor Computing Systems Talos II and Blackbird boards. SMP is not yet supported. Support for POWER8 processors is planned but won't happen unless someone provides us access to suitable hardware.

OpenBSD/powerpc64 does not run under a hypervisor such as PowerVM or PowerKVM.

With regards to b), I am very thankful. My favorite computer is a Tadpole Viper which has a 1.2 GHz UltraSparc IIIi. OpenBSD not only has current support for it, it is also the only thing that will run on it (unless someone has the old Tadpole Solaris 8 or 10 patched CDs). I love that laptop and will be very sorry when it does finally die.

What sort of stuff do you use that machine for? Just curious how people use old (=not new) tech.

I own old computers, from my perspective:

SSH

Writing things

Being productive and having productive workflows in old software, well that's not a sin.

UNIX-y stuff. Nothing I couldn't do with my daily driver, honestly, but the keyboard is better, I like the screen better (1400 x 1050) and I get a dorky smile when I compile things and its sparc64. I just like using it, moreso than the Thinkpad and Acer laptops it has outlasted.

I did have a Tadpole Sparcle (which I liked so much I jumped at the Viper). It came in very handy once when I was learning C and was getting different answers depending on the machine. A deep dive into old newsgroup postings informed me that I had an endian problem. I was excited that I had one of those.

I do have a fondness for computing oddities which is why I bought the Sparcle in the first place and also why I have a "small" 1999 IBM mainframe.

I have an old Sun netra. I occasionally compile software on it just for testing. It's the only real big-endian system I have anymore. It sounds like a jet getting ready to take off, so I don't run it much.

Curious about the "obscure architectures" part. They do seem to have retired quite a few ports...more than have survived. And I don't see anything really outlandishly rare in what's left of the supported list. Maybe the 88k stuff? https://www.openbsd.org/plat.html

I assume the other odd stuff is probably things that are common in China, like the Landisk and Loongson, perhaps because OpenBSD is popular as a router/firewall OS.

Yeah...88k, Alpha and HPPA are pretty obscure for first-tier support being around 30, 15 and 10 years killed by their manufacturer, respectively. Maybe there's still some HPPA boxes grinding out a production workload in some dataccenter someplace. I'd wager to guesss there's some Alphas still running OpenVMS someplace where power is cheap. Almost certainly not an 88k.

Landisk (SH4) and Loongsoon (MIPS), while not common machines, are CPU architectures that are still in production and could be used as a port to newer machines, should someone do the work.

As you have noticed first hand, Theo doesn't "need" to do anything. Neither does anyone else that contributes to open source software. If you don't like it, fork it.

>If you don't like it, fork it.

I'd point out that you don't even need to fork OpenBSD. NetBSD[0], FreeBSD[1] and DragonflyBSD[2] are available for anyone to use, hack on and/or contribute.

This is something of a tangent, but I've discovered over the years that understanding the history of a particular technology (or anything else, for that matter) provides a better understanding of where we are and where we might be going.

As Eugen Weber put it[7] (in a vastly different context, but certainly applicable to this subject): "We're going back to where many of our ancestors came from. To see where their stories came from, and their memories and their habits and the way they are which has made us the way we are. This is what history is about. Where we come from, what lies behind the way we live and act and think. How our religions, our institutions, our laws were made."

I first encountered BSD as SunOS[3] and found it to be useful and workable, although it did have some significant limitations, not least of which was vendor lock-in.

By the time GNU/Linux gained in popularity, most commercial vendors of Unix had moved to SVR3/4[8], and many folks (like this person[4]) chose Linux over BSD for that reason.

In the 1990s, I played with both BSD and Linux, but since Solaris[3], AIX[5] and HP/UX[6] were all SVR(3/4) based and my jobs were to implement/manage such systems, as such I (and many other folks) mostly used Linux personally.

IMHO, that's why Linux initially achieved broader acceptance than the BSDs, even though most folks don't use proprietary *nix anymore.

On a variety of occasions over the years, I implemented BSD variants for specific purposes (generally custom third-party software which required it).

I encourage folks to use the BSDs as well as GNU/Linux when each are appropriate.

Given that all of these OS' are open-source, it seems odd to expend the energy to put any particular OS flavor down, rather than picking one or more that meets one's needs and putting their energy into those instead.

[0]https://www.netbsd.org/

[1]https://www.freebsd.org/

[2]https://www.dragonflybsd.org/

[3]https://en.wikipedia.org/wiki/SunOS

[4]https://rtemsramblings.blogspot.com/2011/11/minix-versus-lin...

[5]https://en.wikipedia.org/wiki/IBM_AIX

[6]https://en.wikipedia.org/wiki/HP-UX

[7]https://www.youtube.com/watch?v=XCyO8meahME

[8]https://en.wikipedia.org/wiki/UNIX_System_V

[Edit: Fixed link references]

Coincidentally, OpenBSD is in fact a fork of NetBSD.

Yep. Precisely because Theo didn't like some of the things NetBSD was doing.

You can always do what Theo did when he was not happy with NetBSD... check out the code, fork, and start your own BSD OS project. The dude is crazy but there is some level of madness that makes projects like these possible. Look at what Terry Davis was able to do, no sane person would be able to do what he did.

Reply on news.ycombinator.com