Linux: Linus Says, Linux Not Designed; It Never Was

http://kerneltrap.org/node/11

by messel • 17 years ago

110 66 17 years ago

kerneltrap.org

messel3• 17 years ago

Pretty fantastic look into how Linus see's technology design, versus evolutionary engineering. F'ing awesome.

Linus' outlook on optimal engineering reminds me of Kevin Kelly's swarm and emergence writings.

blasdel• 17 years ago

Don't tar Linus' acerbic pragmatism with KK's technobabble idealism.

robin_reala• 17 years ago

Probably worth pointing out that this article is from 2002.

mynameishere• 17 years ago

Just a meta-question here: Is pointing out the date of something worth 25 upmods? I mean, when so-and-so got to the forum and saw that...

Probably worth pointing out that this article is from 2002.

...was only upmodded 24 times, did he say to himself:

Only Twenty-Four upmods? This deeply insightful bit of time-tellery is worth Twenty-Five if it's worth a farthing! I shall upmod at once!

s3b• 17 years ago

You're reading too much into it. The points just indicate that around 25 people didn't realize the article was from 2002 until they read the comment so they upmodded

messel3• 17 years ago

I probably didn't deserve a 104 points for finding it, but it's like who's line is it anyway. The points don't matter, what matters is that you guys enjoyed the read.

shachaf• 17 years ago

I like to think that people vote on a comment based on the content, not on the (current) score. It is not particularly unlikely that twenty-four people thought:

This is a worthwhile comment! It points out something that I did not realize, and change the meaning of the article to a degree. I may as well vote on it, that it may move upward in the list and other people read it earlier than other comments.

Also remember that +1 isn't specified to be "Insightful" or "Informative" like Slashdot; nor can you vote a "half-good" comment up a half-point. Instead, to achieve the same effect, a half-good comment must appeal to half as many people. This comment happened to appeal, on its own merits, to twenty-four people.

(Does this mean that the highest-voted comments might be the ones that appeal to the biggest percentage of the site, rather than the ones of the highest quality? Hmm. Does that mean that for the voting system to work well, a good percentage of users must specifically only vote comments that are very good?)

messel3• 17 years ago

It was an oldie, but a goodie. Since I wasn't aware of it (I wonder how much other great info I'm missing), I suspected it was worth sharing. Sun's still kicking though, although more from Java and web growth than from solaris.

diN0bot• 17 years ago

reminds me of yesterday's font page submission on poor use of analogies. not saying it is, but i can't help seeing the connection...oh god! may this is a bad analogy. will it never end!!

Herring• 17 years ago

Sounds interesting. Link?

Not that we really need to analyze it. Theres 50 reasons why this is a bad analogy.

pohl3• 17 years ago

It's interesting that Linus believes that Linux is the most widely deployed Unix. I would have guessed that MacOS X might have surpassed it by now.

Edit: Oops...this is from 2001

yan• 17 years ago

Well, firstly that post was written eight years ago. Secondly, OS X might surpass it on Desktops, but what about servers, computing clusters and so on?

mbrubeck• 17 years ago

Not to mention phones, TiVos, netbooks, routers, Kindles, NAS devices...

pohl3• 17 years ago

Yeah, I missed the date. He didn't need to pile on the caveats and squint to make it true back then.

bad_user3• 17 years ago

Linux is still the most deployed Unix.

Even if a commercial version (like OS X) may surpass it in the future, it will contain parts from modern Linux distributions. And it wouldn't last too long.

What I like about Linux is that it is getting better and better. And it also is the most versatile OS available.

I first used Linux in 1997, when I first tried a Red Hat CD included in a magazine I read. The user experience was just awful, while today's Ubuntu Linux is almost as good as Windows and OS X.

The kernel has also seen major improvements, especially in stability and drivers. My wireless stick, the sound card, my video card ... these work out-of-the-box on Ubuntu.

The communities around Linux may move slower, and have fewer surprises, but they do move forward constantly. In contrast, Microsoft may have the resources to pull off a Windows XP, but every once in a while bad things happen ... like Windows ME, Vista (after 5 or 6 years of waiting), a frozen IExplorer, etc...

One day a distribution like Ubuntu or OpenSuse will be as good as Windows or OS X. When that happens, you can't beat its price or the freedoms that come with it, even if people don't like changing their environment. This is exactly what's happening with Firefox or any other successful open-source projects out there ... once they reach critical mass, they are unstoppable.

RyanMcGreal4• 17 years ago

I had a nice experience last night. I was one of several presenters at a community council meeting, and my netbook - an Acer Aspire One with Ubuntu 9.04 - was the designated AV device.

I was worried and showed up early, but getting my display projected onto the big screen was as simple as enabling the projector as a second monitor and mirroring it with my display. (Also, two PowerPoint shows with animations and so on worked seamlessly on the OpenOffice PowerPoint clone.)

In contrast, I've seen plenty of Windows XP and Mac OSX laptops totally choke at working with a projector.

klaut• 17 years ago

OT: I though OS X was based on BSD and not Linux.

hxa72414• 17 years ago

Natural evolution is best? One word: wheel . . .

Linux as a whole was not designed -- that is easy to accept, and also that it is a strength.

However, design is not so entirely opposed to evolution as might be inferred. Design seems essentially to have some iterative character too. It always entails a some series of ideas, rejections, and then further adjusted ideas. That is deliberately controlled, but it still has variation and selection -- so the possibility of blending between design and evolution is very reasonable.

And given the wheel example, artificial design has something to offer too.

(I think Torvalds' view is consonant with this. -- I mean, is he not overall offering some combination of evolution and design?)

bad_user3• 17 years ago

Your wheel example is not good.

Before wheels, I'm pretty sure people used tree trunks to drag heavy objects from A to B. And I'm pretty sure the wheel's inventor hasn't envisioned the average consumer car we have today either.

Linus's point is that design happens at a micro-level, but at a macroscopic level the direction of a successful project should be made by trial and error with a feedback cycle.

To put things in perspective ... Linux is not designed for servers, it is not designed for desktops, it is not designed for real-time systems, it is not designed for mobile phones. When people wanted to use Linux for such stuff, they took it and modified it to their will, and the parts that slowed them down were eventually removed / refactored.

That's actually very true if you look at the history of many successful projects. And that's evolution more than design. "Release early and often" is exactly about this.

hxa72414• 17 years ago

> Your wheel example is not good.

Well, no it isn't too good. But it does basically fit: It is a small feature of larger systems, as you describe, and it has been done better by artificial design.

gatormax• 17 years ago

Hmmm... idea:

design at the micro-level = hypothesis formulation direction of successful project at macroscopic level = repeated testing and reformulation of hypothesis

in other words, project success is science well applied

Tuna-Fish• 17 years ago

> Natural evolution is best? One word: wheel . . . http://en.wikipedia.org/wiki/Wheel_spider http://en.wikipedia.org/wiki/Hoop_snake

Evolution got there repeatedly. Legs just work better cross-country, and there didn't use to be much roads around.

dstorrs• 17 years ago

Not challenging your point, just curious...do you know of any (non-human) animals that have harnessed fire? I don't mean in an "environmental" way, like looking for food along the edges of a burn, or having IR-sensitive vision, I mean actually collecting it, containing it, using it as a tool. Or is that a uniquely human trait?

hxa72414• 17 years ago

Natural wheels exist, but I wouldn't call them best.

billswift• 17 years ago

Of course natural is best. That's why peregrine falcons are faster than F-16 Falcons.

nollidge2• 17 years ago

Peregrine falcons are probably better at catching food.

eru2• 17 years ago

However F-16 Falcons do not reproduce.

tybris• 17 years ago

Wheels are quite unpractical. Their only advantage is that they require relatively little energy over long distances. However, why not reduce the need for long distances or not create it in the first place?

hxa72414• 17 years ago

Postscriptum:

Saying "wheel!" is crude and invites a simple counter. But the point is there is a simple, important thing where artifice has a substantial lead.

And without well-chosen application of design, the large-scale evolution might be practically excluded from certain areas.

(. . . that was really a side issue. It was the procedural similarity of design and evolution -- they are not so much opposed, as differently balanced.)

RyanMcGreal4• 17 years ago

One of the things I respect about Linus is that he's such a gifted troll. :)

known• 17 years ago

Linus is right here. Kernel is not a Business Requirement. He was scratching his Technical itch.

DougBTX• 17 years ago

The best kind of troll, the one who is correct. Or is that the worst?

btilly3• 17 years ago

It is interesting to read this and compare with Worse is Better: http://www.jwz.org/doc/worse-is-better.html

Note that all of the ways the worse, less designed solution (Unix, C), are better boil down to being better from an adaptation/colonization/survival perspective. In other words the rise of Unix/C is computer evolution in action.

Linus just recognizes that that applies to Linux as well.

ivenkys2• 17 years ago

Absolutely , that was the first thing that came to my mind - this is <i>Worse is Better</i> and taking a further leap its a result of <i>Release Early and Release Often</i>.

cpach• 17 years ago

You can use asterisks to get italicized text: http://news.ycombinator.com/formatdoc

chickamade• 17 years ago

Worse is better is a conscious design choice. Unix/C follow that philosophy but does not mean they are "less designed", but simpler designed.

> Simplicity is the most important consideration in a design.

joss82• 17 years ago

The first Linus comment on that page is the best comment post I have ever read.

It make so much sense ! I can see clearly now !

rudin• 17 years ago

Linus: "Quite frankly, Sun is doomed. And it has nothing to do with their engineering practices or their coding style."

This was in 2001. How does one get such prophetic skills.

rythie2• 17 years ago

I think it was clear even back then sun was the dot in .com and then after the crash everyone started using Linux instead.

dmoney• 17 years ago

Everything is doomed if you wait long enough.

fauigerzigerk• 17 years ago

That's exactly right. On top of that Sun had just lost 80% of its market capitalisation and there was a general sense of doom in the tech industry back then.

jacquesm3• 17 years ago

The Designers of UNIX are called Ritchie and Thompson.

Linux is a re-implementation of that design.

Confusion• 17 years ago

This quote by Linus applies:

But that's like saying that you know that you're going to build a car with four wheels and headlights - it's true, but the real bitch is in the details.

jacquesm3• 17 years ago

Go read an old unix source and tell me if you think it's just the four wheels and the headlights that were common.

Sure there are details, lots of them. Many of those were copied.

And linux didn't even get all of them right, even copying is hard work.

So it's an evolution of the original design.

Especially now, two decades later.

teeja• 17 years ago

That which works well beats that which looks good.

Take quantum for example: When Dirac was asked, "What’s the answer to the measurement problem?" his response was, Quantum mechanics is a provisional theory.

Hexstream• 17 years ago

This surprised me:

"Take TCP for example. The TCP protocol is specified in a series of documents. If you make a formally correct implementation of the base TCP RFC you won't even make connections. Much of the flow control behaviour, the queueing and the detail is learned only by being directly part of the TCP implementing community. You can read all the scientific papers you like, it will not make you a good TCP implementor."

- Alan Cox

tigerthink• 17 years ago

Evolution isn't that impressive. Humans could build human-level AIs if we had millions of years too.

Linux: Linus Says, Linux Not Designed; It Never Was

Related Stories