Slashdot Mirror
Ian Murdock: Linux is a Process, Not a Product
securitas writes "Debian founder Ian Murdock says that Linux is a process, not a product. He also says that the product mentality 'misses the entire point of Linux and the open-source development model.' Because Linux is made up of many different components developed on independent timeframes, Murdock posits, to refer to Linux as a product is to strip it of its dynamism and closes its inherently open nature. Instead, he says that Linux should be viewed as a shared platform and infrastructure technology, and that business models should reflect that or else Linux risks becoming proprietary, closed and just another cookie-cutter piece of software."
Funny, I just got done reading something else that sounds famailar: Science is a process
XML is argued to be a data exchange format, not a data storage format. This article about Linux being a dynamic project has a similarity with XML. It is interesting how people now consider the dynamic nature to be the core of interest, instead of the actual tangible aspects.
I find that the process of creation is only half of the beauty of Linux. The other half is the wide flexibility to use the elements of Linux to do new things. There is no preconcept to the use of the collection of components which give rise to solving an ever widening range of problems and interests
A process is a running piece of code on a UNIX machine. 'kill -9' kills the process, without giving it any chance to recover. It just terminates it right there.
-twb
A product, a process, a community, a method, a team, a concept, an idea, and most importantly, many alternatives.
"Last one in is a rotten goblin!" - Kepp
I think he completely misses some of the good points of having distros unifying the various projects as a unified product but ...
His comments are good because perhaps they can open the business people up to the concept that linux in and of itself is not an OS. It is a kernel with literally thousands of projects built around it to flesh out the total package of the OS.
It is a very hard concept for people accustomed to having their OS as a single product shrink-wrapped and delivered onto them from a single company.
It has its flaws but its a very good article.
For those using linux and for that matter commercial Unix in the IT world, how many bosses actually get the projects as opposed to product conception of Linux?
ACK
Maybe he said GNU/Linux, but the editors changed it. Remember that lots of people have heard of Linux but comparatively few have heard of GNU/Linux. I usually just say "Linux" on Slashdot to avoid people calling me a GNU/hippie and sidetracking the discussion.
who said that security was a process not a product. and that encryption wasn't the ultimate answer (as he wrote for hundreds of pages in Applied Cryptography)...
offtopic yes, but perhaps points to the fact that computer theories, are often in a continual state of improvement and need constant attention.
The majority seem to think that this Linux process can be modified--but only at the cost of losing service and support. The Linux distribution industry needs to start looking at Linux in a new product which will in time, due to the market in technology, culture and process. They need to flesh out the total package of the OS. It is a collection of disparate technologies. Does this mean that the only way to commercialize Linux is an revolution in itself and its progress is already leaving inferior lifeforms (OS) behind. Sincerely, the kernel is the more important part to be distributed to end users as an integrated solution as opposed to just a collection of software components, individually crafted by thousands of projects built around it to make note of it? Reminds me of the underlying platform, putting the user firmly in control of product release timelines and rollout schedules.. Around the edges incompatibilities introduced by the community to come up with a multitude of things besides the simple cake. They think its very important to note that the only way to commercialize Linux is a kernel with literally thousands of projects built around todays leading commercial Linux distributions.
"...or else Linux risks becoming proprietary, closed and just another cookie-cutter piece of software."
Apparently, this guy doesn't understand the process he's writing about.
Weaselmancer
Weaselmancer
rediculous.
I have a lot of friends who use Linux, so I know all the parts about how it's open source, anyone can improve upon it, etc. I think that putting that in the realm of quality would at least cause more businesses to come around. I'm in a manufacturing plant working on outlining our quality system, and it is all about looking at a process for continually making our product better. This is exactly how Linux is made better, not just the kernel, but all the open source software for it. It's like you have a workforce of everybody who uses Linux, and they're all working to make the product better through continual revisions. You mention that to any Quality engineer in manufacturing, you've just sold him on Linux. Yeah, Microsoft releases patchs all the time. These though are coded by what, about 10% of the people out there who improve the code of Linux. This makes Linux far more robust and able to handle different situations. Then add to that the fact that these "patchs" are marketed and priced as a new OS. Once companies, and home users, start realizing this, they'll start converting. The problem is that most people don't know what Linux is, or think you have to have a computer science degree to even use it. Once Linux starts getting away from that image, and people start understanding what kernels and distributions are, that's when we'll see an end to Windows
I've been disappointed by the quality of the linux distributions I've bought in the past. Either my hardware was too new, or using it was just a hassle (mainly due to lack of applications.)
But the thing that got me, in this and other free projects, is how when something doesn't work, it must be the customer's fault. Even if said customer follows all of the how-to's and spends hours compiling (COMPILING!)
Some people put their cars together. Others buy them pre-made. The former group often spends more time in the garage.
Where is quality? Why can't it just work correctly the first time?
I think a lot of people who say he misses the point, are in fact TEHMSELVES missing the point.
Linux, as a collective and generalised OS, is a process.
Debian Linux, RedHat, Slackware, etc. etc., are products. Furthermore, they're comprised of dozens of sub-products, so to speak, each with its own lifespan and schedule.
The general entity called Linux is a procedural entity, or a way of putting together a bunch of products (the kernel, the utilities, the startup scripts, etc.) such that you can make a product with them.
Now this is all fine and dandy. Unfortunately, there are two conflicting results to this:
1) By pushing Linux as a product, you're pushing specific distros which are in effect, proprietary bundles. (Source code notwithstanding, in a professional environment, a bundled distro is _treated_ as a proprietary distro--partly for good reasons!) This is damaging to Linux as a process or concept.
2) Companies don't want to run processes on their computers. They want to buy products.
3) Due to the process nature of Linux, a resulting product (say Debian) is a snapshot in time of all of the subproducts travelling along at different rates. This makes it a big pain for the vendor (and to a lesser extent, the user) to keep current in Linux. This is effectively fallout from Linux "versionitis," and there are no easy ways around it.
"People who do stupid things with hazardous materials often die." -- Jim Davidson on alt.folklore.urban
Linux is...
...)
A kernel
A distribution surname (redhat linux, suse linux yellow dog Linux Gentoo Linux
A trademark owned by Linus Torvalds
A community
A threat to Micro$oft
Free as in Beer
Running my Computer right now
Now it is a process. A rose by any other name...
Are the voices in my head bothering you?
The human body is just a bunch of cell order in such a way that combinds to form one self contained entity. It preforms usefull tasks. Anything that grows, adapts, and replacates can be consitered an orginism. And linux does just that. It is not diffrent than memes. In fact many consiter the open source mentality to be a meme. In my mind, saying Linux is a product is like saying the human race is a single entity. There are thousands of forms and each installation of each form on each computer is diffrent in some way. Why can't I say that Linux is a Species of core software?
sad how many people here don't get it. Ian succeeded in describing what is different about this "linux thing", and one of its major strengths, and many posters here dismiss it with "market-speak". No, sorry, it's much more. Right now we are going thru a new product analysis (hint: initials are BMC) and while I was initially excited that it would run on linux, we find it is only supported for RH7.2 or RH AS2.1. So lame. Instead of this wonderful free, open platform I can modify and optimize, the server turns into just another black box with an expensive (min. $1500) yearly license. Of course at my company, "not supported" is verboten. Very disappointing, and hard to even relate to said company why they shouldn't try to lock it down like every other proprietary platform. These days, we business users are just unpaid (in fact, we pay dearly for it) QC for all the companies we buy broken software from, so locking it down is also preventing us from contributing fixes and improvements.
Thanks again, Ian.
tcboo
This is the best descriptive model I've seen for LINUX, unfortunatly, it ignores the reality of the very real end user. Currently the LINUX end-user is a system-savvy hobbiest or professional, a minimum of one or two levels above the average computer owner (the guy or gal that's still trying to figure out emoticons in AOL-IM or get the time set on their VCR).
This isn't denigrating the average user, it just means most don't lose sleep over the slow adoption of TCP/IP v6. They have little interest in memorizing their monitor refresh rates at various resolutions when DOS (with various windowing programs) and WINDOWS both had easy ways to switch on the fly. Why bother learning the intricacies and simplicity of pipes when all they have to do is hit an icon?
Personally I think they'd be better learning how to work the silly box but the simple fact is this is no longer the era of the ALTAIR,PET/VIC-20/C-64, Apple II, Atari when the purchasers of "home computers" were assumed to have a good basic knowledge OR DESIRED SUCH. Today purchasers just want to get a letter written or look something up on the internet.
Reality says if LINUX is to go further than UNIX did we have to get past the buzz and give the users something more than nine-tenths finished. Patches have to be as easy as wintel or mac machines (forget recompiling, just run the executable) and programs need to be complete and usable as delivered, including example templates, complete help files (written in ENGLISH!), and even online help (ala the much hated though immortal clippy).
The issue is not the developers, where LINUX is now strongest. The issue is the "mom & pop" end user that want's another toaster. Steve Jobs understood that with the original Mac, Bill Gates still does, the question is when will Debian, Red Hat, Suse, etc. catch on.
It doesn't matter what you wrap your emotions around, Reality is a brick wall specifically designed to scramble eggs
This is what I thought when I saw the headline. You are of course correct, I have a thing running my computer (a product), not a way of doing things.
;).
While I have my objections to the semantic game that he's playing trying to redefine the word Linux, (and his strange business-ease definitions of the word product) I do understand what he's saying. Essentially he's trying to differentiate Linux from other software products. If it were me I'd avoid the semantic game, and say that Linux is a product that requires a different environment and attitude than your traditional OS (for both the software makers, and end users). Of course that's just not quite as sexy as saying "linux is a process". A headline like that wouldn't get me on the front page of Slashdot
That is the great strenght of Linux and GNU you are perfectly free to use it as a dev platform in anyway you see fit. If your version is worth while then it will survive. It is the perfect dev platform for really advanced embeded systems. The ideal thing is to create a killer device which becomes a real product. At that point the software becomes secondary, and if you need to bow to Redmond or where ever so that your system can work then you are at a disadvantage when the company whose OS you use decides that your device is something they really have to own. Patent devices not software. Give the really inventive people freedom from rediculous constraints. GNU/Linux is the way of the future. Let MS patent every concievable system function software sequence and
and eventually new American tech will grind to a halt.
Look how long affordable 64 bit systems have taken to reach the market. This is purely the doing of IBM, Intel and Microsoft. But then again what does a home user need 64 bit for. Of course there are no applications that a MS wants to think of. The RIAA would have kittens if 24/96 recording became easy on the home computer. Also small art schools would be able to do too much. Budding digital artists using Maya and like tools would get too good too quickly. Advanced scientific tools available to all schools and teachers.
Oh hell you cannot have little people doing things that only rich guys can do.
This is the reason why high tech is going off shore, not that we are stupid just that we are stupid enought to let the major corporations control the future of tech. The real cost of advancements in computer tech has been the software. GNU/Linux has thrown a wrench in the works and eventually will open up 64 bit tech in the Orient and Europe. This will happen so fast that Microsoft, IBM, and Intel will not even know what happened. American government intervention on their behalf (like what happened with tron) will not help the giants this time. Compete or die is going to be the answer from the government in future. As so it should be.
OH THE SHAME I fell off the wagon and use sigs again!
Linux does not live or die by GNU alone. It is a system, and yes, GNU is part of that system
Wrong. GNU is an operating system. It's a replacement of the proprietary unix operating system. GNU was meant to be a system, Linux was not. Linux is a kernel, which needs an operating system to live. Theoretically (and perhaps even practically) Linux can run with some other operating system. But it usually doesn't.
It seems rather childish for RMS to stand up and shout "Well, if it weren't for ME you wouldn't even have your ball to play with!"
RMS doesn't say that. RMS says he's the one who made the operating system. And the operating system really should be called GNU. Not GNU/Linux, just GNU. The main reason RMS suggests to call it GNU/Linux anyway is that the name "Linux" is something that people know. It's pretty bad PR to give yourself a completely new name when people are just getting to know the thing. That's why he wants us to call it GNU/Linux. So people still recognize the word. If many people call it GNU/Linux, perhaps we could start to call it by its proper name.
In a later post, you say that Linus didn't try to finish the GNU system, he tried to make his own project instead. I don't really know about Linus' reasons, but let's suppose that this is true. That's fine, the nature of free software is that such things are possible. But it is usually considered polite to use the name of a system that the creator chose for it. If Linus didn't want to call the whole system "GNU", because he thought his contribution was large, then he should have thought of calling it something like "GNU/Linux". But to leave out the name of the original is quite rude, really.
I have to agree with rknop. Debian Stable is pretty static, but there are those moving targets, such as the unstable branch (which can be a hell to setup, when mixing several sources), or the other distro's, based on Debian (to name 2: Knoppix and Libranet (my favorite)).
Never caught the purpose of this 'HURD' thingy though. Isn't that some hardcore Stallman stuff?
So "used" cases that used "unused" could break, though older compilers in essence used "unused" to mean both "used" and
In Korea a person tried making it his own trademark.
a Boston man tried taking control of it.
Microsoft Purchased it
although it's currently controlled by The Linux Mark Institute so Linus doesn't have to deal with it.
Never caught the purpose of this 'HURD' thingy though. Isn't that some hardcore Stallman stuff?
I'm sure a lot of people thought the same about Linux when it started. It does provide the same functionality as other Unix like systems, but it is based on a different design philosophy. It is completly non-monolithic
Almost everything is running in userspace as a server, (except for the microkernel), it promises greater scalability through its massively multithreaded and highly granular design.
Yeah, it is Stallman's concept and part of his project, but it is different enough from everything else out there to be significant.
Check out the Debian Gnu/Hurd for a distribution, and the GNU Hurd page for some docs.
Read, L
Like Ian Murdock says, Linux distributions are comprised of many different software packages, all contributing to a working whole. I think this is why Gentoo is doing so well.
I'll be the first to admit that Gentoo has been difficult to set up. Because I'm not a distro maker, my box doesn't have the slick polish of Lindows or Red Hat. On the other hand, now that I am set up, Portage (Gentoo's package manager) is so flexible, powerful, and up-to-date that all of the software I care about is as fresh as CVS and the software I don't care about is just not installed.
To me, it doesn't even make sense to put a version number on Gentoo, because it is always changing, and always current.