Slashdot Mirror
The Benefits Of Radiation On Linux
roblimo pointed out this article in Business 2.0 by one Clay Shirky, a professor of New Media at Hunter College. Shirky goes into the idea of "adaptive radiation," and makes some pointed observations about its effect on software, Linux specifically. An excerpt: "This will keep Linux out of the best of breed competition because it is never perfectly tailored to any particular environment, but it also means that Linux avoids the best of breed trap. For any given purpose, best of breed products are either ideal or useless." Shirky is an insightful commentator as well; he also set up a "Web Pay Phone" outside his office at Hunter a few years back. Now go mutate!
I didn't read this article as anti-Linux in any way. What's anti-Linux, or anti-Free software, about it?
More the opposite -- yes, he makes the claim (I disagree) that Linux "isn't ready for the desktop," but it appears to be because he is expecting some pretty great things out of Linux, not because he dislikes it in any way -- but on the whole, he makes the argument that Free software is often / usually the best thing *overall* even if there's some place for non-Free.
Windows doesn't have the same sort of freedom to radiate.
I guess we read this totally different ways.
timothy
jrnl: http://tinyurl.com/c2l8yr / foes: http://tinyurl.com/ckjno5
I think this sort of capability is best provided by the GNU Hurd.
TO BUY A NEW CAR WOULD MAKE YOU SEXUALLY ATTRACTIVE.
He was just BSing his way thru a performance review. I would guess that it went something like this:
Reviewer#1: Looks like this Clay Shirky guy spends all day chatting on AIM, downloading MP3s on Napster, and surfing porn sites. We will have to fire him.
Reviewer#2: Mr. Shirky, can you give us a good reason for your behavior?
Shirky: Ummmm....It's is because....I want to...be...a professor of New Media? Yeah, that's it.
Reviewer#2: Is that a position we have?
Reviewer#1: Well, if he lets us help surf the porn sites, we could make one.
Shirky: Deal!
No replies made to AC posts. Please log in.
I can't speak to #1, but as for #2, you work on the net for several years, found and sell a business, and then teach Web design and internet theory in a college that uses slightly outmoded jargon.
In any case, I am on leave from Hunter, but the New Media program goes on apace.
-clay
Slashdot is essentially a committee of 250,000 usernicks, and by its very nature as a committee, it will generate results inconsistent with past statements and declarations. Slashdot is enormous compared to its size one year ago, much less two or three years ago, and the influx of new users alone would account for the change in pary line (or at least commanding philosophy).
With that said, I have to pick your point about "One License, one GUI, one Platform". To be sure, "One GUI" and "One Platform" are counterproductive, but "One Liscense" has its benefits. Some of the biggest problems and fights in the open-source movement have come from the selection of different and incompatible liscenses -- a defacto standard (like the GPL or LGPL) would ensure that if your efforts on your neat project and my efforts on my neat project don't have to be mutually exclusive or incompatible.
I'll even go so far as to say we should stop arguing about forks and mutations altogether. As long as the source (and therefore the structure) is disclosed, any incompatibility can be overcome within a single generation. It's just not a problem.
-- Anne Marie
Try, an old technique and still the most effective on it's own. Radio therapy is still more effective than chemo! Combined they are even better. Radio therapy may not work for all cancers, but where it's appropriate it's your best chance. Look up Mecical Physics some time.
Friends don't help friends install M$ junk.
I guess I would argue that the Linux distribution exhibits "punctuated equilibrium" as opposed to "adaptive radiation", especially considering the relatively recent advent of it's global distribution and acceptance. Ontology recapitulates phylogeny.
I know all analogies have breaking points, and I'm about to step over this one's ;)
Which would you rather have as a pet?
Mutts tend to be better behaved, healthier and smarter than the blue-bloods, at least when it comes to dogs.
Yeah, he's got a barb about Linux / Free software being unable to be "best of breed" ok, but it depends on what your context is to determine what "best" means. If your requirements include future flexibility and lack of royalties etc, then it could easily be the best.
timothy
jrnl: http://tinyurl.com/c2l8yr / foes: http://tinyurl.com/ckjno5
His second point is obviously incorrect: if Linux on x86-based PCs attracts more coder-hours of work than Microsoft can hire for Windows, Windows will eventually be enough worse than Linux that buying Windows will be an obvious bad idea, not just the bad-idea-you-can-gloss-over it is now. At that point, Windows' market share will start dropping.
His first point MAY be incorrect, and he presents no evidence that it is not. Just because billions of Linux variants are being created for new platforms doesn't mean that the core portions of Linux itself, the ones that don't change from one platform to another, are being improved. For example, I suspect that the majority of these ports are doing absolutely nothing for the window managers. Also, there is the question of whether any changes at all are leaking back from these efforts into the larger community; if the people doing a port don't bother to tell anybody about the changes they've made in a useful way, none of their work benefits Linux itself.
Hmm, Wouldn't Radiation and Linux produce those "Mutants" That Microsoft was talking about in their anti-Linux Ad :)
--
In Soviet Russia, Trojan exploits YOU!
Restated:
"It means humans, while versatile, will never be the best at anything because of the way we are built and designed. By extension, this also applies to ANY organism which can survive outside of specialized niches: primates makes likeable apes instead of Bengal tigers."
I may be nuts, but I think it's very possible for a "likeable mutt" to wind up being the best overall, far and away, because these "likeable mutts" have no real Achilles' heels. The Bengal tiger is nearly extinct; human beings are still undergoing population explosion.
As we have seen in many biological environments, evolution does NOT happen at a continous rate. Instead, species tend to evolve very significantly in a very short amount of time, and then stay pretty much the same untill this happens again. This theory is known as Punctuated Equilibruum, and is currently accepted in the world of biology. WHen mapped to software, each brief occurance of rapid evolution can be thought of as a software release.
Throughout History, these punctures usually result in an extremely rapid power shift. (Think of the chimps when Homo Erectus first showed up.) In the end, that species that wins is the most able to adapt. In other words, if this evolutionary model works (which it sure does seem to!) than the new release of a distro might cause the dominant OS to change extremely rapidly. Further, given the change in useability between Linux PPC 2.5 years ago, and Linux PPC right now, I think we are certainly on the right track.
Although the author of this article would surely disagree, Linux has the potential and will, become the dominant desktop (as well as server, and every other market) in the extremely near future. So don't give up!
Ok, I'll stop rambling now,
--Alex the Gnome Fish
But we tend to forget that the atom-blaster points both ways.
Well, if I am still living in the same dimension I was 5 minutes ago then there is something real wrong 'round here. /. was mostly a pro-standartisation... If my head is still in its place I still remember posts talking about how good was RedHat and how bad were the kernel mutators... Well time seems in place. I don't see nothing strange 'round here. Is this and aftershock of /. getting hot about his journalistic quest? Maybe... A 180 degree move and not quite for the right place... Too much 180 degree.
Well frankly I am a mutator partisan. Paraphrasing Zhirinovskii, the reactionary political radical schizo here: "Mutate linux. Mutate the progs, the scripts, the source. Mutate the kernel! Everything should mutate, even the Tux the penguin and Torvalds." Maybe this is tooo radical. But still sounds good. As what makes linux is its capabilty to mutate. Think about what would happen if we sticked to "One License, one GUI, one Platform". That is exactly what M$ does. And that is what most companies do willingly or unwillingly.
If Linux had stucked into a strict environement then it would be never what it is now. Linux is the most multiplatform of all systems, from 86 to S/390. Meanwhile it is the systems that carries most of the Old and New World. In fact it is not only a continuation of UNIX but also of DOS and old Windows platforms. It even carries enough power to carry/emulate software from platforms that are long forgotten, such as ZX Spectrum or Commodore. Or to hold up emulation of weird platforms like Nintendo or Palm. It is a system capable of being a server or a desktop system. In this last point I should note that the author of the article is deeply wrong. linux is already ready for many desktops. I have seen/participated on the implementation of desktop systems for Internet & Office appliances. 68% of people don't want Windows back. And only a few stubbornly decry Linux as an Office system.
In one point I would still declare Linux a drawback. On vector design systems. Here Linux is still a way to go, even if Corel 9 is already out. On the rest Linux is already conquering. Yes, it is HARD to install it. To professionaly and carefully install it. But it is a system that works on the base "fire & forget". Once fired, technical support becomes miserable. So technicians are kicked to R&D, making its progress even faster and smoother.
It is mutation that makes this. And this is what M$ should be afraid of. Because this system is capable of inveding every computer and make it much more faster and reliable for a wide range of tasks.
However this should be taken with a few grains of pepper. Mutations are good if we can control them. If we change things too much then we may face future incompatibilities. But this does not mean keeping the kernel in one piece or flaming someone for choosing BlackBox against Gnome or KDE. The problem are not the changes but the rules of conduct to implement them. some sort of comitemments that would allow every program/mutation/patch/fix/implementation to live peacefully side by side. and to be used when they are needed, where they are needed. This is much better than having an abstract "Unity" on Linux.
While stray EM radiation seems to crash my Windows machine often, my Linux machine seems to put up with EM radiation. I can't say I've ever noticed a benefit of EM radiation on Linux....
As far as generalizing this to all Open Source project, though, I'd have to disagree. It really depends on how focused the project leader is and how much 'me too' syndrome there is hovering around it.
If the project is fairly low-key, or the developers are strict, the project will get the features it needs and people itching to implement something unnecessary will either not know about the project in the first place or go find something else to do. (or maybe even make a fork)
With more lenient project leaders, it's much more succeptible to "feeping creaturism". Instead of becoming specialized for one area, it gets the kitchen sink.
I'm not saying either method is bad, just that it is possible for open source to produce the "blue bloods". I'm a code optimization / embedded systems junkie, so I like to use code size as an example. IIRC, Linux was first coded on a 386, then as the hardware got faster and the features creeped in, Linux grew. It still works on the older machines, but probably not as well as it used to. XFree86 is another of those quite spread-out projects. It works on diskless terminals, big SMP machines, and it can even be crammed into an embedded system like the iPaq. But, a GUI specially designed for small systems will always be better at running on small systems and there's no reason it can't be open source too.
Ack! I've been rambling!
-- 2 + 2 = 5, for very large values of 2
*ahem*
NT slams linux at SMP. Believe it. Though I'm sure that won't be the case forever.
Seriously.. I'd like to see the NT kernel open sourced.. as well as NTFS. They *are* good, and have *great* potential. Microsoft just keeps making these godawful operating systems with them!
I mean, originaly, when NT was called the 'unix killer'... from a kernel perspective, it was quite true. NT is superior.. but MS keeps building these shit OS things with it.
Tux grows to 40 times original size, stomps off to terrorize Redmond.
More radiation and he'll gain the power of speech: "Hello, Mr. Gates..."
- M
-----
Go ahead, blame me... I voted for Nader!
Casting Bill Gates in the role of senator Kelly? See the diatrabes against the mutants! See the kidnapping and exposure to radiation! See him covered in slime on a beach in New Jersey! Oh wait, all the swimmers on New Jersey beaches look like that...
I'm trying to teach myself to set people on fire with my mind... Is it hot in here?
> Linux is still a server platform and not a desktop platform, it doesn't try to be everything.
I don't think you can say that Linux was designed as a server OS. After all, Linus started it for his own desktop.
What the deal is, Linux just tries to be an operating system. You can layer server software or desktop software on top of it. Or both.
And that, IMO, is what is allowing Linux to "radiate" into virtually every computational niche known to humaniti.
Sheesh, evil *and* a jerk. -- Jade
Is it really better to be second best in several areas than to be number one in a few areas?
Yes. Think of it like the OS Olympics, where Linux grabs a silver in fifteen events, while Windows, Solaris and BeOS each get five gold medals. I'd rather go home knowing I was the best all-around athlete than knowing I could only throw a discus farther than anyone else.
.sig: Now legally binding!
It means Linux, while versatile, will never be the best at anything because of the way it is built and designed. By extension, this also applies to ANY Open Source project: Open Source makes likeable mutts instead of blue bloods.
Note how the effect of the GPL is on these embedded devices and the use of IP in embedded devices was not even mentioned.
If the GPL remains toothless, then these issues won't matter. If RMS and FSF ever get around to suing some people WRT the GPL, then the embedded world will rush to BSD.
(Richie can't win....if RMS sues as he should be, then no one will use the GPL. If RMS doesn't sue, people will use the GPL but RMS sells out on his principles.)
The missive you reference was written by Tim Cavanaugh. Tim does a bunch of writing for Suck magazine, which is one of the few consistently good web 'zines out there (and in which, there ran an entire article about what Tim refers to as scare quotes. Don't know if he wrote it, though.)
I may be nuts, but I think it's very possible for a "likeable mutt" to wind up being the best overall, far and away, because these "likeable mutts" have no real Achilles' heels. The Bengal tiger is nearly extinct; human beings are still undergoing population explosion.
From a genetics standpoint you are correct. I used to work at a company that delt in selective breeding. While working on one project, I was looking at the genetics charts and came to a sudden realization. The perfect animal is the perfectly average animal. This was quickly confirmed to me by several of the geneticists.
You see, genetics is a bit like a balancing scale. You can't make something physically stronger and then expect that it won't be at a disadvantage in say its immune system. Or make it prettier, but not expect that its feet and legs will get softer. This is how genetics works. In order to create an animal that is as perfect as possible in all areas, it must be completely average. Of course no such thing exists in nature or really in captivity because the environment affects 80% of the animal's genetic outcome.
Javascript + Nintendo DSi = DSiCade
Free software is both infinitly configurable and arguably best. Anyone can take a plausable solution and make it into exactly what they or their organization needs. The result is better for everyone and forms the next plausible solution.
Friends don't help friends install M$ junk.
2) How does one become a "professor of New Media"? I guess I'm just surprised to see academes recognize a movement [outside their own].
--
Anthropology has long said much the same thing about humans: as general-purpose creatures we never will be the fastest, have the sharpest claws, climb the fastest, run the fastest, or have the quickest reflexes. We win by not being the best, but by being the most general-purpose and therefore the most adapative.