Linus' Other Gift to the World

Glyn Moody writes "Linus is widely recognised for initiating two major developments: Linux and Git (it's an interesting discussion which of the two in the long term will be regarded as more important). But there's a third, which people tend to overlook: he also pioneered the key ideas behind what later came to be called open innovation. As more and more companies open up to embrace customer-generated ideas, and the idea spreads to other areas like open government, perhaps it's time to add open innovation to the list of Linus' achievements."

Comments

I love git!!

[Anonymous Coward]

Git is super amazing!!

Re:I love git!!

[Anonymous Coward]

All the SVN tards on here will disagree with you, no doubt.

Re:I love git!!

[Anonymous Coward]

I think most here would prefer Git. SVN is far more corporate.

Re:I love git!!

[Necroman]

I think the better comment is "DCVS is super amazing!!"

Many people forget that there ware 2 other decent implementations of distributed source control out there (Mercurial and Bazaar), both of which function rather closely to Git. Though, from what I've seen, Git is currently the fastest and most efficient when it comes to processing various commands, but they all do everything rather quickly.

Git I would say is popular in the open source world for 2 reasons: Linus uses it and Github. My gripes still stand with it requiring Cygwin on Windows and its weird terminology (which is backwards from many of its predecessors). It's a great tool and I'm happy that its pushing people away from CVS and SVN, but it's not perfect and it's not the only DVCS on the block.

Re:I love git!!

[mikael_j]

I find Git conceptually interesting but it took me a while to grasp the beauty of it, mainly because everyone I knew who I asked to explain it to me would point me to various "explanations" of Git which were huge swaths of text and graphs that described all the nitty-gritty technical details but skipped basic "big picture" explanations of it. So I pretty much came away from it knowing as much as I did before reading the "guides", that it was some form of decentralized VCS...

Eventually I took it upon myself to read up on it a bit more and began to see the usefulness of it although I have found an alarming number of guides and explanations of Git that deal with the question of how to set up your own "central" Git repository for all your code (pretty popular with geeks who, like myself, have a home server with redundant storage and regular backups) with "You're not supposed to do that! AAAGH YOU JUST DON'T GET IT!!11".

Re:I love git!!

[bberens]

I don't have anything against DVCS but I personally haven't switched from SVN. SVN doesn't cause me (personally) any headaches that might be solved by DVCS. For some, I can see how it's quite useful and beneficial but I suppose you could count me in the group that doesn't prefer it.

Re:I love git!!

[DrXym]

Git is great but the Windows tools are still a bit poor with msysgit & TortoiseGit being about 95% the way but still a bit flakey compared to TortoiseSVN. Fortunately for Java development, EGit is coming along very nicely and supposedly part of Eclipse 3.7.

Re:I love git!!

[Anonymous Coward]

Fast, cheap topic branches. Git knocks svn out of the park in the regard, even if you never use the distributed part.

Re:I love git!!

[Anonymous Coward]

I didn't much care for git when I used it from the command line. But with a nice gui (yeah, I'd rather click to stage a file than type it out), I find it generally better than svn. Branching in svn is such a pain that I didn't do it. git branching is painless. I still prefer having a central repo. My major problem is that git pushes mmap to the limit and exposes some mmap bugs in debian/arm/linux, so it's unusable on my server.

Re:I love git!!

[Anonymous Coward]

You say that, as if it were a bad thing.
The more killer features Linux has, that Windows hasn't, the better.
Look how Compiz alone made hot girls want to install Linux!
HOT GIRLS!
Seriously, I alone know 3 who installed Linux because of it.
HOT... GIRLS!!
That's like ... the be all, end all argument. Like a cat pushing a watermelon out of the ocean. Like OVER 9000!
HOT GIIIIIIRLS... WANTING LINUX!!! HOLY SHIT! ;)

Re:I love git!!

[tom17]

So i'm still getting my hands dirty with git, having not really used a code versioning system before. I have come across a lot of information of the type that you describe so am still in the "some form of decentralized VCS" camp. I do not yet see the big picture.

What is it for you that made you see the picture, and can you explain it easily like the sites I have seen can't?

Also, I too want a 'central' git repository and I too have seen the "You're not supposed to do that!" but I still don't get why. I want to be able to sync the project on any computer i'm working on and be up to date, surely centralised is the way to go?

Re:I love git!!

[Anonymous Coward]

I was raised on CVS and then switched to SVN. Going over to git was weird at first, and it felt awkward. I kept at it though because of some of the nice features, and then one day the light clicked "on." Using CVS or SVN managing branches and tagged releases becomes very burdensome quickly. They also require you to work out the management of your repository with your team in great detail. Then when someone forgets and does something differently it can make a big mess.

With a tool like git *everything* is a branch. You just branch branch branch, and git has some very nice tools to bring branches together in a way that is very comfortable once you get used to the idea. The thing that is great about git is that it manages things with respect to the branch and the source. That way each person can manage their own repository the way they want. The system also allows different subgroups to share their work without using the main repository.

When it comes time to bring everything together you can choose which branches to push and you can choose which repository to push it up to. The result is that you have a huge amount of flexibility in how you manage commits, and a wide range of approaches can be used. git lets you decide how to handle the different branches whereas SVN dictates the way you approach the management of your repository.

In terms of branching, now days I use git for my personal projects. Even if I do not want to share my repository the branching tools make for a very natural work flow. Once you get over the idea of "branch == bad" it really frees you up to play and try things without fear of things going bad or getting your repository all clogged up.

Re:I love git!!

[tom17]

For me, it's just little personal projects that I am working on. No collaboration at the moment and no branching. I just want to plod along on my little projects, keep the versioning controlled incase I make a big old f-up and have it available wherever I am (At home on my laptop, at home on the desktop, in the workshop, on the train, at work). Whenever I work in a new place, I just want to continue where I left off in the last place.

Is git the wrong tool for the job?

I understand about the branching and don't have anything against it as I don't have any pre-learned version control methodologies in my head already (other than lots of messy commenting and file renaming lol) as I have not used a cvs before. I figured from my initial research that git seems to be the way forward and that I should learn and use that rather than another.

Re:I love git!!

[jd]

SVN has all kinds of issues. It really needs a therapist at this point. Git is the best of the open-source systems. Of the closed systems, Perforce and Plastic are the only serious contenders, but I've had a hell of a time with Plastic due to the way it identifies individuals, problems with unmerging and database organization that's kludgy at best. Perforce is ancient, has some amazingly bad quirks, and is more expensive than hell above a couple of users.

I've burned many brain cycles trying to decide what to do, since Git just doesn't have the kind of front-end I'd like but does have a nicely-designed engine. At present, it's like driving a Volkswagon Beetle with a Formula 1 engine - the power is amazing, the capacity is staggering, the risk of ploughing off the road and over a cliff unrivalled. Everything else is like driving a Fornula 1 car with a Volkswagon Beetle engine - safe in the same way other statues are.

Re:I love git!!

[DrXym]

Windows support is still a little unpolished but these days you don't need Cygwin. You install msysgit which is natively compiled and comes with a bunch of Unix tools for running scripts and so on. Then you install TortoiseGit which invokes msysgit and presents things in dialogs for pushing, pulling, cloning, diffs etc. It works much the same way as TortoiseSVN.

There is also Eclipse Git which is implemented over JGit. So if you develop in Eclipse it's basically what you would expect of any source control system and very nicely integrated. Eclipse is moving to Git for development so it's likely to gain a lot more prominence in Java land.

Re:I love git!!

[Anonymous Coward]

If you just want to keep a version history then anything will do. RCS will do that and works fine. That is how I used to view version control software. Since switching to git I have started using it as a tool to help develop and not just keep track of file histories. It has become second nature to just branch, play, test, and then if things worked out then merge back in the appropriate place. It makes it very easy to keep multiple versions and experiment with different approaches. git is more than just a tool to keep track of history. The flexibility and ease of handling branches helps make it a development tool and changes the way you approach a task.

Re:I love git!!

[vbraga]

Is git the wrong tool for the job?

No, not at all. It still the right tool. You still have less overhead than if you were using SVN, since you no longer need a repo server and such. But you will only really learn git when you need to collaborate with someone else, this is where it shines.

The only thing I miss from SVN is TortoiseSVN, which is a really polished and nice GUI. TortoiseGit is not even close (when I'm using Windows). On Linux I just stick to the CLI out of habit.

Re:I love git!!

[Anonymous Coward]

that's an advantage of git over svn, not an advantage of DVCS over non-DVCS.

Re:I love git!!

[icebraining]

Eventually I took it upon myself to read up on it a bit more and began to see the usefulness of it although I have found an alarming number of guides and explanations of Git that deal with the question of how to set up your own "central" Git repository for all your code (pretty popular with geeks who, like myself, have a home server with redundant storage and regular backups) with "You're not supposed to do that! AAAGH YOU JUST DON'T GET IT!!11".

You can do that, but you'll be mixing commits from different projects in the same log, which is messy, for no good reason. You'll also have to sync all the repository even if you just want to get one project's files.

Unlike with SVN, Git repositories are extremely cheap (git init .), and backup is just copying the directory, so why not create one repo per project?

Re:I love git!!

[thetoadwarrior]

Mercurial ftw.

Re:I love git!!

[Anonymous Coward]

bazaar is horribly slow. At least 2 orders of magnitude slower than git.

Re:I love git!!

[serviscope_minor]

I think the better comment is "DCVS is super amazing!!"

Yes it is very handy.

Many people forget that there ware 2 other decent implementations of distributed source control out there (Mercurial and Bazaar)

Not to mention darcs and monotone. I've never used montone, but darcs is very nice, except that it probably doesn't scale as far as git and doesn't like very large files.

I'm happy that its pushing people away from CVS and SVN

To be honest, in general, VCS is super amazing. Even CVS works well enough for small projects that it's often not worth switching.

Re:I love git!!

[xophos]

You forgot one important point: Git is FAST. With larger Projects the other two sometimes take some time to think.

Re:I love git!!

[diego.viola]

+1 for Git, working with it is simply amazing.

Re:I love git!!

[Dr_Barnowl]

I quite like Bazaar. Not as blazing-fast or powerful as git, but does have the luxury of both simpler userland tools, and a better SVN integration plugin (git-svn has some nasty gotchas, bzr-svn has fewer).

Re:I love git!!

[Dr_Barnowl]

All the DVCS systems I've used do topic branches better than SVN. SVN branches cheaply, just like the new kids, but it's merging sucks.

I tried some very simple refactoring tests on SVN recently, aware that it had received some love in the "merge" department, and it fell at the first hurdle - renaming files. It just isn't suitable for any kind of modern development method.

Re:I love git!!

[oursland]

SVN does NOT branch cheaply. You end up with duplication of all the branched code. DVCSes like Git do not duplicate, but rather create a new pointer to the same code. Only the changes have to be stored in git.

Re:I love git!!

[Opiyum]

I don't have anything against DVCS but I personally haven't switched from SVN.

I hear you... but after using Git for a few months at work for internal development using github; I DO have something against DVCS and Git. It is fast, and I recognize that it addresses a different use case than most VCS's (the D) and has advantages for distributed teams who may or may not ever merge. I also recognize continuous code integration suffers greatly and push's tend to be rare... separating (fragmenting) code base into multiple repo's disconnects the history and code relationship of your source tree, and basically, Git pushes agile dev teams away from each other and imposes many issues with code layout, etc. With SVN I can just check out the specific code I'm interested in, make my changes, commit and I'm done. Very helpful if your repo is a gig or more ... and before you chime in, VCS's repo's should be large ... if they are not then you most likely are not using your VCS to it's advantages ... or ... maybe are just really unproductive.

I've been trying to find a place to say this for a few weeks now, thank you for giving me the opportunity!

Re:I love git!!

[polymeris]

fossil, fuck yeah!

Re:I love git!!

[Dr_Barnowl]

An SVN repository is a filesystem with an additional dimension (revision) - all branching does (or any copying) is copy a pointer to an inode. Branching in SVN is just as cheap as branching in git ; they both do the same thing, storing a pointer to the top node in a tree.

If you don't believe me, you can try it ; branching a large tree in SVN only results in the growth of the repository by a few KB. You can look at the size of the revision, as it's stored in a numbered file of it's own.

Where it falls down, is as I said, on it's merging, which was never designed to cope with the idea that branching was a frequent or trivial activity, as it is with DVCS systems.

WTF

[BlackPignouf]

Is this a new : trend to add colons in the middle of a sentence?

Re:WTF

[TheRealFixer]

use Open::Innovation;

Re:WTF

[Anonymous Coward]

Is this a new : trend to add colons in the middle of a sentence?

should only be used to extend the sentence

Re:WTF

[sgt scrub]

Agreed. That is much more classy.

Re:WTF

[Anonymous Coward]

I think this is the new global::Innovation();

Re:WTF

[LordLucless]

No, it's also generally used to indicate the start of a list.

Re:WTF

[jd]

Depends. If we assume "later" is a time-traveling host and Open Innovation is the open port, it's ok. :: would imply that OI is IPv6-compatible.

Re:WTF

[Anonymous Coward]

Is this a new : trend to add colons in the middle of a sentence?

In fact, colons should only be added in the middle of a sentence.

Ummm no

[Anonymous Coward]

Because Linus simply did not do that.

FFS, just post a 'missed connections' in craigslist if you love the guy that deeply, don't waste space here on it.

Re:Ummm no

[Samantha Wright]

Exactly my feelings. "Now, Mr. Submitter, is that GNU/Innovation that Mr. Torvalds gave us, or BSD/Innovation?"

Re:Ummm no

[Magic5Ball]

What a completely forgetful article.

I'd give the cookie for coining "open innovation" to von Hippel (for making it popular, regardless of who "invented" it) but the concept and practice has been around for decades.

Fairchild Semiconductor (and many other electronics concerns) practiced open innovation in the 1960s by involving customers in the design, manufacture, and specification of their products, providing royalty-free reusable circuit and implementation designs and libraries, hosting user groups and forums, etc. (See Brock, 2010's "Makers of the Microchip", MIT press.)

Before semiconductors, agricultural machinists in the 19th century shared comparable information and enabled users to collaborate to increase user satisfaction and thereby customer loyalty, etc.

The only news here is that the software, the most portable and easy to replicate of modern industrial technologies, took decades to adopt open innovation as mostly a token gesture. But that's only if you believe the article and ignore all the microcomputer zines and technical bulletins starting from the 50s in which computer operators from IBM and big institutions down to the basement delay line memory hacker openly shared code and implementation details with each other.

Speaking of tokens, take a guess at what year the first two digits of IEEE 802.x designate.

Post-singularity enterprises

[mangu]

I think in the long run people will have forgotten both Linux and git, but the open enterprise system will go on.

In the future when we will have abundant robotic power, corporations will have to be managed differently. People with managing ability today are people who are good at manipulating people, with automated systems managers must be people who are good at manipulating machines, i.e. programmers.

The catch is that programmers aren't very good at manipulating people, and that include their peers. In a typical enterprise today a lot of effort is put into negotiating between the different departments and divisions. I cannot imagine a company managed by programmers doing that.

The Linux management system will work when managers no longer have people beneath them.

Re:Post-singularity enterprises

[Anonymous Coward]

Open source governance is likely to be the most revolutionary, game-changing thing to happen to society since the French Revolution.

Re:Post-singularity enterprises

[Noughmad]

I'm still waiting for it to happen, and I'm not holding my breath.

Re:Post-singularity enterprises

[WaywardGeek]

I wish that were true, but I fear open source governance to date has gotten entirely away from the sort of innovative leadership Linus provided. Let's take Debian, for example. Innovation back when everyone compiled everything from source, and it was truly a bazaar model, was off the charts. Then came the Debian package manager and Debian official commiters, and multi-year release cycles, all driven with open source governance. What we have now is a cathedral model, where a minority of the community determine what goes in and what doesn't, and where any new innovation takes years to get into the stable release. What Linus provided with the Linux project was solid leadership in a bazaar model that no longer exists, except for those of us still willing to download and compile from source.
I find, now that I'm developing for Windows and Android, that innovation is more rapid there, because you can write code and have it in user's hands without the red tape. No good governing system should create barriers between innovators and users, and potential super-star leaders need to be empowered to lead. Linus ran Linux development in a way that connected users and innovators, while keeping final authority over his project in his hands. A good bazaar needs a skilled benevolent dictator, not a innovation crushing exclusive republic like Debian.

Re:Post-singularity enterprises

[DrgnDancer]

He's not talking about governance of Open Source software groups, but rather applying Open Source concepts to government. It's a clever idea that largely ignores all sorts of political realities. The Wikipedia page he links will describe it better than I ever could.

On the subject of your comment, I find that I agree with you, but I'm not really sure how else it can go. Anytime a software package, whether an OS or application, becomes popular enough to have a significant percentage of its users be non-programmers it's going to lose its "everyone contributes to the whole" vibe. Debian is the way it is because the vast majority of Debian users either can't or don't want to compile everything on their systems from source. Unless you're going to require everyone user to be a geek and a contributor, you're going to eventually get an elite of coders who are willing and able to contribute directly; and various levels of users below that that have steadily less of an idea what's going on behind the scenes.

Re:Post-singularity enterprises

[WaywardGeek]

I would say that history so far proves your right. However, I feel we can do better. For example, I'd love to see an open-source cross-platform app store for Windows, Mac, Android, Linux, and iOS that enables coders to get to a lot of users with minimal red tape. I like the Android model of running apps in jails with detailed level of user controlled permissions, and the way Android never upgrades a dependency from under your app to cause it to crash. It simplifies package creation and reduces the need for long term user testing before exposing an app to average users. I also like the Ubuntu launchpad services which automate building packages for multiple platforms from one source. User PPA's are one of the most important ways right now that developers can innovate and reach users quickly. I love the Debian model for sharing libraries where I can search through tens of thousands of existing libraries with one command, and install the one I want with one more. If we could combine the best of these systems in new open-source cross platform super-highway between developers and users, we might be able to get open source innovation back on track.

Re:Post-singularity enterprises

[Anonymous Coward]

I think in the long run people will have forgotten both Linux and git, but the open enterprise system will go on.

In the future when we will have abundant robotic power, corporations will have to be managed differently. People with managing ability today are people who are good at manipulating people, with automated systems managers must be people who are good at manipulating machines, i.e. programmers.

The catch is that programmers aren't very good at manipulating people, and that include their peers. In a typical enterprise today a lot of effort is put into negotiating between the different departments and divisions. I cannot imagine a company managed by programmers doing that.

The Linux management system will work when managers no longer have people beneath them.

So, what you are saying is: in the future, there will be robots?

Re:Post-singularity enterprises

[Anonymous Coward]

Well, there's your problem right there. How's it going to take to happen if your entire contribution is to wait for it?

If you are waiting for your benevolent politicians to bestow it upon you, you will be waiting forever. (And going backward all the time.)

If instead you start coding or otherwise contributing, you can actually make it happen.

Re:Post-singularity enterprises

[Anonymous Coward]

Well-well look. I already told you: I deal with the god damn customers so the engineers don't have to. I have people skills; I am good at dealing with people. Can't you understand that? What the hell is wrong with you people?

innovation.

[bobs666]

I thought innovation was what the Big Corporations did after the patent's ran out.

Take the case of the X-Y-Box (the first mouse) it was patented in the 60's. and low and behold on the 80's we got GUI's with mice. makes one think. Is this kind innovation setting us back 20 or 30 years.

Re:innovation.

[Anonymous Coward]

No. Considering that it only sets us back 20 or 30 years would be to ignore the negative synergistic effect on innovation from the incarceration of any given technology by the patent system.

Re:innovation.

[Anonymous Coward]

A patent does not just set us back just 20 or 30 years, each piece of technology imprisoned in a patent has a negative effect on what gets invented afterwards. The cumulative effect goes far beyond the damage each individual patent has - negative synergy.

Re:innovation.

[Anonymous Coward]

I disagree. In 1960, AIDS was patented. 20 years later, it goes out of patent and suddenly you can't buttfuck a man wearing leather in a bathroom without getting sick. We need longer patent durations to prevent these tragedies from happening.

Re:innovation.

[Anonymous Coward]

Not a good example. "We" got mice the moment 128k computers that could run mice became cheap & abundant. Corporations aren't at all shy about licensing tech to make consumer products -- take a look at cell phones if you want to skip all the examples in desktops and laptops.

I'm not saying you can't make a case about patents hurting innovation, just not here with this.

Filthy Athenians!

[Anonymous Coward]

They're always stealing Linus's ideas before he gets to them.

Old Idea + "do it online" != innovation

[Anonymous Coward]

Nothing against Linus, but I think it's a little silly to take a normal idea, apply "do it online", and then call it a new idea.

Collaboration has been around long before Linus. Perhaps Linus is one of the first to collaborate online... ok, that's great. Comparing this "accomplishment" along side Linux and Git is a little silly. Linux and Git are genuine accomplishments which Linus should be very proud of.

Re:Old Idea + "do it online" != innovation

[panikfan]

Nothing against Linus, but I think it's a little silly to take a normal idea, apply "do it online", and then call it a new idea.

Collaboration has been around long before Linus. Perhaps Linus is one of the first to collaborate online... ok, that's great. Comparing this "accomplishment" along side Linux and Git is a little silly. Linux and Git are genuine accomplishments which Linus should be very proud of.

Agreed, was thinking the same thing when I saw this story...

Just maybe,

[d3m0nCr4t]

in a few centuries it will be regarded as the begining of true democracy! :)

Re:Just maybe,

[Anonymous Coward]

In a few centuries the world will be united under Sharia law!

I think we just have it labelled wrong here

[hellfire]

Linus' major achievement was popularizing and demonstrating open source and the projects it could accomplish, and Linux and Git were merely demonstrations of that. Glyn merely has caught up to us who have realized sometimes great inventors great invent things, but in software great inventors truly only invent great ideas. That is what Linus has done here. Stop thinking of Linux as a thing and start thinking of it as an idea, part of a greater idea which he has touted for a very very long time.

Re:I think we just have it labelled wrong here

[Anonymous Coward]

If what is great about Linux is the idea, then RMS deserves the credit.
But I reject the premise. What is great about Linux is the implementation.

Re:I think we just have it labelled wrong here

[denis-The-menace]

I thought it was RMS, too.

That being the case, Linus' real "other gift" is providing proof to the world that RMS' idea was valid and possible.

Re:I think we just have it labelled wrong here

[WaywardGeek]

RMS since 1984, when he declared the GNU free software operating system effort, has been the GNU/FSF champion. He invented the GPL license, which along with similar licenses that have been developed, are the legal cornerstone enabling software innovators around the world to collaborate on free software efforts. However, with the emacs project, RMS showed a lack of skill in encouraging innovators to go nuts with whatever development they like. Everything had to be approved to get into the official emacs release, and RMS was a single point of failure, constricting the flow of innovation from others while providing most of it himself.

Linus, on the other hand, has the people skills needed to enable innovators to contribute while standing enough out of the way to keep from restricting innovation. That's why he's credited for this model of governance. That's the model we need to figure out how to replicate. Unfortunately, it seems to require a benevolent dictator who is brilliant, has thick skin, and good people skills. If we can figure out how to replicate that kind of success, based on how Linus did it, then we can credit him for the model, but so far it seems that Linus' model requires a very rare kind of benevolent dictator who is brilliant technically, has good people skills, and likes to enable others to innovate as much as they like to contribute themselves.

No good model for sharing the efforts of software innovators exists today. Of all platforms, Android seems to have the most innovation at present, simply because they allow any old coding fool to publish an app with minimal red tape. However, Android more or less forbids sharing between developers. We'll never get where we need to be with the current models.

Re:I think we just have it labelled wrong here

[DrXym]

That being the case, Linus' real "other gift" is providing proof to the world that RMS' idea was valid and possible.

It's valid and possible if you put pragmatism before politics. This explains why Linux is everywhere and Hurd isn't.

Re:I think we just have it labelled wrong here

[he-sk]

From what I read, the output of Stallman and Torwalds is roughly comparable. They're both awesome programmers who have the ability to create well-designed code in a short amount of time.

However, whereas Stallman started out with a clear agenda, an extreme sense of perseverance and tried to build a free Unix system no what, Torwalds basically got lucky. If there hadn't been a nearly-complete GNU system (written by Stallman and others) at the time he wrote his terminal emulator^W^Wkernel, Linux would very likely not have taken off.

PS: Why does Slashdot not support <strike>? Oh that's right, Slashdot sux.

Re:I think we just have it labelled wrong here

[Xtifr]

Not just RMS (and the GNU project), but to some extent, the BSD, X11, TeX and Perl projects as well, as long as we're talking historically. As for implementation, Linus is a clean coder and good at organizing people, and I admire him for both attributes, but the overall implementation of the Linux kernel has a lot of warts, IMO. Regressions are far too common. If it weren't for the much broader hardware support in Linux, I would probably prefer to run a GNU/BSD system. (I may still if Debian GNU/kFreeBSD progresses well.)

I think it's quite possible that if not for the legal battles back when Linux was still little more than an interesting toy, BSD would be the dominant free OS today.

I don't use Perl much any more, but in its day, it was amazingly innovative and was used in a boggling variety of places. It was often described as "the duct tape that holds the Internet together" for good reason.

If we take a loose definition of "open" (as distinct from Free/Libre), I might have to throw James Gosling (xemacs, Java, and much more) on the list as well.

Re:I think we just have it labelled wrong here

[LWATCDR]

Linus popularized it. Before Linus if you asked the average person what free software was they would have told you it was.
a. A copy from a friend.
b. That public domain junk like lemon aid stand.
c. What I can download from the Pirate Chest BBS.

I actually remember hearing about GNU Unix way back then. I was so excited and could hardly wait for it to show up. Even giving RMS the credit for coming up with the idea of free software is a big stretch. People had been passing around the source for programs for a very long time. Things like the games like Adventure, Star Trek, Lunar Lander and Rouge are prime examples. What he did was come up with the GPL and then sprinkle it was an extremest philosophy that all software must be free or we are all slave to the man. RMS did write EMACS and GCC which are vital FOSS to this day well at least GCC is. But even giving RMS all the credit for FOSS is short changing all the history of programmers sharing code that happened for decades before and thinks like BSD. I will say that they are right on the money with statement that Linus popularized FOSS. If he hadn't I think people would have long since forgotten the GPL and the GNU project and we would all be arguing over FreeBSD vs OpenBSD, and NetBSD.

Re:I think we just have it labelled wrong here

[Anonymous Coward]

I thought Git was a "I had to write it because my preferred closed source version control tool is no longer free-as-in-beer for me, due to those pesky pirates and telnet terrorists".

Or was BitKeeper just a dream I had?

Re:I think we just have it labelled wrong here

[Anonymous Coward]

Oh, bullshit. Hurd isn't everywhere because it was technically terribly hard to implement, and interest in developing it waned when the Linux kernel came along.

Re:I think we just have it labelled wrong here

[DrXym]

That's an incredibly weak excuse. If Hurd had been goal oriented and pragmatic it is likely it would be what Linux is today.

Re:I think we just have it labelled wrong here

[Urkki]

Linus, on the other hand, has the people skills needed to enable innovators to contribute while standing enough out of the way to keep from restricting innovation. That's why he's credited for this model of governance. That's the model we need to figure out how to replicate. Unfortunately, it seems to require a benevolent dictator who is brilliant, has thick skin, and good people skills. If we can figure out how to replicate that kind of success, based on how Linus did it, then we can credit him for the model, but so far it seems that Linus' model requires a very rare kind of benevolent dictator who is brilliant technically, has good people skills, and likes to enable others to innovate as much as they like to contribute themselves.

Biotechnology will soon solve the issue. We have our selection of BDFLs in open source community. All we need to do is create hybrid "clones" of them, raise them as nerds, then a few decades later see which did best, and repeat the process. It won't take many generations until we have the ultimate BDFL at our disposal.

Re:I think we just have it labelled wrong here

[Urkki]

I think it's quite possible that if not for the legal battles back when Linux was still little more than an interesting toy, BSD would be the dominant free OS today.

I believe the reason Linux rules over BSD is license. BSD license allows companies to "just take", and in short term this is so much more tempting, that rate of "giving back" isn't as high as it could be. There are manager level people in companies using BSD software, who abhor the idea of giving away company secrets built on top of BSD licenesed technology, even when they're not really secrets. They're, perhaps justifiably, afraid the their competitors will just "take" and not reciprocate, and gain advantage.

And then there are more than a few developers, who really don't like a company "taking" their code the way BSD license allows.

With GPL, this is different, and I suspect with operating system and especially it's kernel, the difference gives GPL a clear advandatage. It's got nothing to do with luck, and all to do with basic human feature: greed, which manifests in unwillingness to give without having guarantees of others doing the same.

It's interesting how Apache has been very successful with different license, but I just wonder if Apache software would be better today if it had Affero GPL style license, having gained many many improvementes from everybody using Apache software, or if it would be long dead...

Re:I think we just have it labelled wrong here

[Anonymous Coward]

I know we'd all like to think that everything is easy if you really try, but no.

  Having talked to some of the folks who've worked on it, it's a fucking bear to get right. They haven't tossed out several designs because they're all la-di-da flower-sniffing hippies who don't have a clue where they're going, it's because SHIT. COULDN'T. BE. FIXED. The initial design plan was far too optimistic about the ease of designing a microkernel, and they've paid for that mistake, badly.

Re:I think we just have it labelled wrong here

[DrXym]

Well that's the point. They were trying to be perfectionists, not pragmatists and got nowhere fast. If they set aside some of their more ambitious goals they would have made more headway. In doing so Hurd would have become useful for something, attracted more users and some of those users may have been motivated to fix the things that were bypassed on the first iteration.

That's why Linux succeeded. The Linux 0.01 kernel made a beeline straight for a functioning x86 kernel and fleshed out from there. People joined the project because where there was no way to run a Unix like OS on x86 for nothing before, now there was. Stuff like multiple architecture support, networking, filesystems etc. came later when critical mass was achieved for development. Pragmatism made it successful. This is highlighted in other ways the project has proceeded from choice of source control to the licence it uses.

Re:I think we just have it labelled wrong here

[lucian1900]

About benevolent dictators, they don't all require a thick skin. Guido v.R. is a good example of a "softie" BDFL.

Re:I think we just have it labelled wrong here

[Anonymous Coward]

Being "perfectionists"?? Jeez, I'm sorry. I didn't realize you were dense!

  The HURD kernel was a very difficult project right from the get-go, and no "pragmatism" could possibly be applied when microkernels, hailed as the wave of the future by people in the know back then, were not yet well understood. You can't take an easy road when you're travelling through unmapped territory, and there was no way to know at the outset that microkernels would turn out to be as tricky as they are. When it was begun, everyone "knew" that micrkernels were the way things would be done from now on, and if GNU didn't use one, it would be obsolete out of the gate. Maybe they should have accepted that the great microkernel transition would be slow and rocky, and just used a macrokernel with increasingly microkernel-like features, the way Linus would do, but that would have required a fucking crystal ball, not "pragmatism."

  The pragmatic view at the time was "microkernels are superior in every way, and everybody's going to be using them. We have some of the brightest programmers around, building one out of this example kernel should be a snap!"

  You're trying to apply hindsight to the past.

Re:I think we just have it labelled wrong here

[DrXym]

Stop being so absurd. Seriously. You're trying to make out that a moribund dead project which accomplished fucking nothing in 20 years is pragmatic? Hiding behind a microkernel is no excuse for the godawful mess of nothing that was Hurd. I was programming QNX around the same time, a real time OS that used a microkernel that was knocked together by a relatively small company. It wasn't the only microkernel either with Coherent & Minix being other examples. If a few guys could manage to produce a dist in the absence of tools like gcc, bash etc., Hurd sure as hell had no excuse.

Um no.

[Bill_the_Engineer]

Linus Torvalds has done well with Linux and now Git, but I don't really seeing him deserving the title of pioneer of "open innovation". At least not in the way the author is using the word "pioneer". Linux being the most popular open source project makes the project itself the catalyst for this so called "open innovation". There is a difference between taking advantage of open source methodologies and creating methodologies. The author has seems to lost the grasp of that difference in his zeal for idol worship.

I think RMS deserves the title of the creator (or pioneer) of "open innovation", and that says a lot since I don't always agree with his philosophy. RMS is the one that really stuck his neck out and preached the gospel. Even today he is either loved or hated by the software community.

Then there are all the advocates that came and went during the lifespan of Linux. They wrote manuscripts, sold the idea to their employer, or invested their own money in open source development. During all this time Linus focused on his vision of the kernel, and having a take it or leave it attitude towards advocacy. His main concern was making a quality kernel and rightly so.

If we used the word pioneer correctly then we would consider Linus one of many pioneers in this open source crusade. There are a lot of them with the scars from all the arrows in their backs. Sorry for the slight negative tone, but the idol worship in that article irritated me this early in the morning. More coffee!

Re:Um no.

[Anonymous Coward]

I agree that RMS deserves a lot of recognition for his early vision and good social conscience, but if you look at how key pieces of the GNU system was developed, it seems rather "cathedral-like", to borrow the over-used ESR analogy; even today it seems more so than the Linux kernel. FSF carries a powerful social message, but the early days could also be understood in terms of the clubby hacker cliques that I imagine RMS participating in at MIT. I don't see it as very open to outside contributions. Consider for example the Emacs schism of the 90s, for example. If it were simply about the code, RMS would not have maintained such ridiculous positions on the "official" Emacs 19 code base. Similar can be said for GNU HURD -- if they had been more open about its early development instead of taking a "we know better than you, you'll wait 'til it's ready" approach, maybe it would have had a higher chance of taking off.

Re:Um no.

[sgt scrub]

I would agree with you but RMS contributes more philosophically where Linus contributes more physically. For example, Louis and Clark were pioneers. They didn't come up with the whole manifest destiny "god said we could have the land if we explore it" attitude. The concept came from some politician. That isn't to say I would equate open innovation with manifest destiny.

Re:Um no.

[antifoidulus]

Thats the one thing that always gets me about RMS, for someone who ostensibly espouses freedom, he doesn't handle dissent very well..... "You are free as long as you do what I tell you to do"

Re:Um no.

[Dog-Cow]

I guess you didn't read the article, or your comprehension is very poor, as he very clearly states the differences between what Linus and RMS did. And he didn't say that Linus pioneered open development or open/free software. He says that Linus pioneered the approach of asking potential users and developers for feature requests, while RMS wanted people to work on his stuff.

Re:Um no.

[cerberusss]

but the idol worship in that article irritated me this early in the morning

Thank you for your wisdom, oh wise one!

Re:Um no.

[staalmannen]

Well, in fact Linus Torvalds was used as counter example to FSF in the "Cathedral and the Bazaar". Most open source projects prior to Linux was either horribly fragmented or horribly centralized by requiring copyright assignment to FSF or some other organization. In that way the manner in which Linus let others contribute to and influence Linux, he actually demonstrated another model of development (which inspired to the "Cathedral and Bazaar" text, read it if you haven't). People seem to think that Linus is a celebrity by choice - from what I have experienced, he is rather uninterested in those things. I think there is a big difference between idol worship and admiration/respect to/for someone who does something really well. Since Linus is not trying to sell us something like ideology (RMS) or shiny trinkets (Jobs), I tink he is worthy of some admiration. http://www.catb.org/~esr/writings/cathedral-bazaar/

Re:Um no.

[Bill_the_Engineer]

I did read the article and didn't buy into his comparison of the two. The top down structure versus asking for volunteers on the internet over simplifies a lot. For instance, RMS has been doing this much longer and part of that time he couldn't take advantage of the internet. When I first started using GNU software my first access to the products was via floppy disks through the mail. Creating a unix like kernel using floppy disk or BBS distribution is significantly harder than downloading from USENET or kernel.org

Re:Um no.

[Bill_the_Engineer]

What about gcc and emacs? They both existed longer than Linux and enjoy a much larger install base.

I don't really understand your point. You seem to have completely disregarded the fact that not only did RMS delivered "physical" software, he did it before Linus.

Re:Um no.

[dkleinsc]

RMS doesn't exactly deserve that title either.

What RMS was doing was exactly what most of the mathematical and scientific community was doing already, namely sharing ideas like crazy and giving everybody who knew what they were doing a way to contribute. And it had been what was going on in Unix-land before it became heavily commercial, most notably in BSD Unix.

As much as RMS is portrayed as a radical, the real radical position is the idea that software source code should be hidden from users and protected from prying user eyes for the purposes of profit.

Re:Um no.

[TuringTest]

No no no you don't understand... people are free to do what they want, and RMS is free to criticize them for it.

Re:Um no.

[Anonymous Coward]

And despite having made great advances such as making source available and coming up with GPL, he did so with a very closed development model that was at times needlessly hostile to outside contributors. Google "lucid emacs" for a representative sample of where this went wrong.

Re:Um no.

[sgt scrub]

The word "more" is important in my post. And yes, without gcc most FOSS applications wouldn't exist including the linux kernel and git. However, I was pointing out that Stallman's contributions were not just physical but philosophical. Also that his philosophical contributions were "more" as in greater than his physical contributions. To simply say Stallman contributed gcc and emacs in comparison to Linus' kernel and git would leave out the philosophical contribution by Stallman. Thus. One was "more" physical. The other was "more" philosophical.

Re:Um no.

[Anonymous Coward]

Exactly. I see Richard Stallman and Bill Gates as being two sides of the same coin.

Re:Um no.

[swillden]

However, I was pointing out that Stallman's contributions were not just physical but philosophical.

Umm, the GP's point was that Stallman wrote GCC and GNU EMACS. How is that not a physical contribution?

Re:Um no.

[Bill_the_Engineer]

What RMS was doing was exactly what most of the mathematical and scientific community was doing already, namely sharing ideas like crazy and giving everybody who knew what they were doing a way to contribute.

That is not my experience with the mathematical and scientific community. It is true they are open and peered review, but they are definitely not that way from the very beginning. It's very competitive and they hold on to key pieces of data that they worked on so they can be the first to publish. It's not until after they publish their findings when they start freely sharing ideas like crazy.

Now compare this to open software development where people don't have to show their doctoral credentials before they are allowed to contribute. Also the goals of the software are known from the very beginning and additions can be submitted from anyone from anywhere. The key difference being that it is a community effort even prior to first release.

The two communities may look similar, but they are definitely not alike.

And it had been what was going on in Unix-land before it became heavily commercial, most notably in BSD Unix.

I do agree that BSD and public domain software was sharing software longer than FSF. However the distinction I was giving RMS credit for what is considered the open development model where all of the derivatives remain open and available to original author and the community. BSD and public domain do not have this trait. BSD Unix has an active open development community but it's derivatives can and are being used in commercial and closed source versions. It is the principle behind GPL that allowed Linux to garner the popularity that it enjoys now even eclipsing that of BSD.

I agree that it would be more accurate to call RMS a pioneer of the all software shall remain open movement.

As much as RMS is portrayed as a radical, the real radical position is the idea that software source code should be hidden from users and protected from prying user eyes for the purposes of profit.

I would have to respectfully disagree with you on this point. Considering the major contributor to the origins of computing in the US was the DoD, the idea of software code being secret isn't really that radical. That doesn't even take into consideration the competitive nature between DoD contractors during the war effort (WW II not today :P).

Re:Um no.

[Anonymous Coward]

It's also not a new phenomenon. It's been going on for centuries.

Read "Collective Invention" by Robert C. Allen.

http://p2pfoundation.net/Collective_Invention

Re:Um no.

[Anonymous Coward]

He says that Linus pioneered the approach of asking potential users and developers for feature requests, while RMS wanted people to work on his stuff.

That just shows how poor the author's memory is. Plenty of projects had user feature contributions before Linux - UNIX and Spice come to mind. What would happen was the companies that owned the software would appropriate the user contributions and incorporate them in their closed source projects. That's the whole reason the GPL exists and created a model where contributors don't have to fear their work being monetized.

Re:Um no.

[Bill_the_Engineer]

That may be true. Not all pioneers got it right the first time. In fact I think the number of pioneers who did is rather small or practically non-existent.

Re:Um no.

[DrXym]

egcs is another example where people got frustrated with the gcc development process being too slow, too dictatorial, too strung out and forked the thing. In that instance egcs was so successfull that it became the basis for gcc 3.0 demonstrating forks don't have to stay forked forever.

Re:Um no.

[Anonymous Coward]

People who start their posts with the "um" faggotry shoul be taken out back and shot. Douchebags.

Re:Um no.

[Bill_the_Engineer]

Understood.

Let me use a sort of car analogy, it's Slashdot after all and I haven't made a terrible one in a while.

There was a need to have a paved road to connect the east coast with the west coast of the US. Before the road existed, groups of early pioneers travelled westward for a better life and many of them were never heard from again. It took a bold pioneer to not only see the need for a paved road but make an effort to build the roads and bridges over rough terrain. He had to sell the feasibility of the idea to potential supporters, the government and builders. Because of that bold pioneer and the people that shared his idea, other people can now take advantage of the paved road to do more than anyone thought possible.

So I considered RMS the bold pioneer who had a desire that all software should be free. He preached his gospel to potential supporters. He built emacs and gcc that could be used to develop the free software. He did this over rough terrain (no world wide web at the time).

I consider Linus another pioneer who's job wasn't as hard due to all the groundwork laid by RMS. He only had to concentrate on building Linux. Truth be told there was a symbiotic relationship between RMS and Linus Torvalds. Linus wanted a free unix compatible kernel and needed help creating it. RMS needed a catalyst to propel his philosophy into the view of mainstream geeks. FSF saw the potential for Linux and used it as an opportunity to advocate for free software. Linux benefited from the advocacy due to the increased number of volunteers. Let's not forget the others that advocated on the behalf of the free software movement and made it more palatable to mainstream businesses. So my point all along was that there were more than one pioneer in the "open innovation" movement. However if I was asked to name one person, I would have to say RMS. Like I said in my first post in this conversation, this is despite the fact that I don't always agree with RMS' philosophies.

Re:Um no.

[Jeek Elemental]

Ill second this, I respect Linus as an engineer first and foremost. This is ofcourse extended to all of the kernel hackers.
Technical advice from any of these carries weight.

RMS I respect as a visionary with quite extraordinary consistency and patience, again extended to all GNUs.
He gets a lot of flak from short-sighted people for an uncompromising stance, this is unfair.
Its his role to be clear and firm, so that stuff moves in the right direction instead of stagnating in details.

They have overlapping spheres of influence, I think a large part of free softwares success is due to them both working at the same time.

Re:Um no.

[Urkki]

That's the whole reason the GPL exists and created a model where contributors don't have to fear their work being monetized.

To nitpick (or perhaps not), GPL does not prevent others from monetizing contributions, nor does it try to do this, at all. It's just more difficult while adhering to GPL requirements, and requires a less direct approach than selling draconian licenses to use the software. But some approaches not only work, but also scale very very well, for instance Red Hat's approach.

Re:Um no.

[Anonymous Coward]

"I would agree with you but RMS contributes more philosophically where Linus contributes more physically"

Uh-huh. RMS invented the concepts, encoded the philosophies, setup the organisations, built the unix clone... and linus added a crappy, monolithic kernel, using RMS's license and toolchain, and failed to get or acknowledge the true philosophical intent of what RMS was doing.

Linus is awesome. Git is mindblowingly cool, especially given the ramifications now that github etc. are around. But RMS, he is not.

Re:Um no.

[Kuruk]

He did start something truly good for humans.

Sharing with everyone is a basic good move.

Governments and Corps can spin it as much as they like.

There stuck with open source. It will get bigger.

lunix

[eyenot]

it tupid

Linux vs Git?

[dr_tube]

Can someone explain, or point to a discussion, of how it is argued that Git could be more important in the long term than Linux? Isn't Git small fish compared to Linux?

Re:Linux vs Git?

[royallthefourth]

Linux is an OS kernel. Kernels come and go over the years; Linux has been around for a long time already. Eventually, its development will hit a wall created by some ancient design oversight and someone else will make something new to resolve it.

Version control systems have more staying power. The popular SVN was developed as a better CVS, which dates back to 1986 (according to Wikipedia), which itself was an improvement over something else from the 70's. Linux is roughly analogous to SVN's place in the chain of operating systems: it filled a broader need, but it wasn't really something totally new. Git came along and changed the whole version control landscape with its distributed model, since I guess it was the first DVCS to also be open source (although maybe Mercurial is a little older).

If Linux hadn't been invented, someone else would've created a UNIX-like system for the PC. We probably would be using BSD instead, and most of us wouldn't know the difference. Git, on the other hand, was the innovation that most people didn't even know they wanted.

Re:Linux vs Git?

[doti]

There is nothing special about Linux, apart from being a successful FLOSS operating system.
There were no great genius work from Linus there, he "just" wrote an ordinary OS and made the source public.

Git, in the other hand, is a work of a genius. It's not merely one more version control system.
It made branching, merging and atomic commits so acessible that it changed the way people code.

ps: no idea why you were modded down, it was a completely reasonable question.

Re:Linux vs Git?

[dr_tube]

OK, I guess I never had a feeling for why there haven't been more MS competitors out there when writing a good OS is not such a big deal. As an outsider, it always seemed like writing linux must have been a bigger deal than maybe it was.

ps: A long long time ago I somehow obtained terrible karma during a period when I never posted (I've never figured out why), and I guess I don't post enough to raise it back up.

Re:Linux vs Git?

[Anonymous Coward]

Eventually, its development will hit a wall created by some ancient design oversight and someone else will make something new to resolve it.

Hmmm, no. The code will simply be rewritten to remove said design oversight and Linux will keep on chugging. It's happened dozes of times over the last 20 years and therein lies the beauty and power of Linux, it's fluid and dynamic, it evolves and adapts. Why do you think Linux became so popular? Because everyone can tinker with it and if your idea is actually good enough, it will get integrated.

Re:Linux vs Git?

[Xtifr]

It's not merely one more version control system. It made branching, merging and atomic commits so acessible that it changed the way people code.

Are you sure? I never used Bitkeeper, but I know that Git was heavily inspired by it. Also, while I admit that Git is an incomparable improvement over SVN, I'm not convinced its as much of an improvement over other DVCSes like Arch, Bzr or Mercurial. Mercurial in particular is a very close match in features and flexibility (though I admit to a mild preference for Git myself), and is arguably more accessible (in the sense of easy to learn and get up-to-speed with) than Git.

Re:Linux vs Git?

[tbird20d]

Oh for crying out loud. Linus is smart, but git started as just an opensource rehash of bitkeeper. Should we credit Larry McVoy with genius?

Re:Linux vs Git?

[Anonymous Coward]

I'm sorry, but Git has a horrible interface, and doesn't do anything that wasn't already available (in open source.) As far as I can tell, it was designed to make people feel 1337 when they manage to figure out how to do what should be a trivial operation. Git is, IMNSHO

Re:Linux vs Git?

[iggymanz]

not special? getting so many companies and people to *contribute* is the outstanding thing about Linux compared to the others. Why do the open source alternatives lag in device support? (and yes, I prefer some of the alternatives to Linux but still that mindshare thing is formidable)

Re:Linux vs Git?

[ceswiedler]

It's ridiculous to assert that Git is anywhere near as important as Linux just because "version control systems have more staying power" and "someone else would've created a UNIX-like system for the PC". Git is quite nice but its impact is miniscule compared to Linux. Moreover, the revolutionary thing about Linux was its development model (which BSD didn't have) which is precisely what led to those revolutionary ideas in Git.

Re:Linux vs Git?

[Urkki]

Git is quite nice but its impact is miniscule compared to Linux.

Git allows places like gitorious.org, github.org. While seemingly like eg. sourceforge, I think git makes a huge difference here. It's a game changer (which is kinda orthogonal to being "nice" or not).

Re:Linux vs Git?

[Anonymous Coward]

> he "just" wrote an ordinary OS and made the source public.

Except that he didn't write the OS. He wrote a kernel and ported userland from other projects that existed because of Unix. Torvalds has done everything in his power to avoid the crap he didn't want to care about- userland- and in return we got the situation "Linux" is currently embroiled in (severe fragmentation).

An "OS" is more like Haiku or OS X or Windows. You get a complete package. With Linux, you get a kernel. Even if you just want a simple shell, you have to go to a third party to download and compile one.

-AC

Re:Linux vs Git?

[Anonymous Coward]

W-w-wwait - did you say people started coding differently, beacause of git? What more nonsence can you say?
Git is a simple filesystem based CVS which takes the idea of storing everything in a folder/file based way, because the linux is written this way. Git came after linux to help linux get better.

Re:Linux vs Git?

[ceswiedler]

You are massively understating the impact Linux has had on the computing world.

Re:Linux vs Git?

[Urkki]

Oh I agree Linux has significantly bigger impact, especially today. But impact of git is just happening. And I think git will "win" over eg. Mercurial for a few reasons. First of all there's modularity vs. monolitic, which allows git to have much wider range of uses, and also allows more directions for open development, enabling innovation, inviting innovators, increasing developer base, speeding development. The impact doesn't have to be big to accumulate over the years. Then of course there's the... for lack of better word, brand: "Made by Linus". It influences choices some people make, again growing user and developer base. For these reasons I believe git has more enduring future than eg. Mercurial: it will be(come) the default choice for those who specifically don't want to use it for whatever reasons.

I'd hazard to say, git will be a choice much like Linux is now, even in companies. There are alternatives, but... why bother with them, "everybody" knows Linux/git, it's the safe choice for some things (much like the old saying "nobody ever got fired for selecting Oracle database").

Of course with version control systems, it's often trivial to switch from one to the other, "installed base" has less meaning than with operating systems, and new contenders have much better chance of being accepted to wider use. This evident from different version control hosting sites often allow a choice of VC to use, still even including plain old CVS.

Re:Linux vs Git?

[Anonymous Coward]

> It made branching, merging and atomic commits so acessible that it changed the way people code.

Er, maybe it did for users running Linux or using Github, but I can assure you that things weren't quite as nice for Windows Git users trying to limp along with the official Cygwin-based build of Git for Windows and eGit without involving Github at some point along the way about 9 months ago. I made the mistake of agreeing to move our project from Subversion to Git at the behest of a new collaborator last summer, and progress basically ground to a halt for 3 months until I decided enough was enough and moved everything back to Subversion. Compared to Subversive (Eclipse), Netbeans' built-in Subversion support, and SyncroSVN, eGit and the official Windows Git client were positively stone age by comparison.

Now, I'm sure the situation will be much better in another 2 or 3 years, but we were *totally* unprepared for the crude state of Git in Windows-land circa September 2010. The glowing online articles singing Gits praise were (in retrospect) almost ALL written by Github users. Remove Github from the equation, and the experience of using Git under Windows quickly went down the toilet... then clogged it, overflowed, and left you stepping around a minefield of poo.

Git per se might be brilliant, and might very well be the wave of the future... but for Windows users who don't want to pay for Github, the word "future" is very, very important and needs to be emphasized.

Re:Linux vs Git?

[Anonymous Coward]

Git built upon BitKeeper, which built upon TeamWare. The genius belongs to others, not Linus.

Do not comment on something you know nothing about.

Rock on

[Blakey Rat]

Let's claim he won WWII, too. What the hey. It's not like anybody will actually check.

/ Godwin

Re:Rock on

[0100010001010011]

Lets claim that Hitler was killed by a bad Git Merge.

THAT is a godwin.

As an online discussion grows longer, the probability of a comparison involving Nazis or Hitler approaches 1 (100%).

Re:Rock on

[garaged]

As much as I like/love Linus, you got a great (funny) point, wish i had mod points

Re:Rock on

[Splab]

Git is the Chuck Norris of 2011!

Slashdot and Twitter

[cosmas_c]

Cosmas_c is widely recognised for initiating two major developments: Slashdot and
Twitter ... (comments follow?!)

perhaps it's time to add bad karma to the list of Cosmas_c' achievements.

Why credit Linus

[kbw]

I remember when DDJ published the BSD i386 source, that was pre-Linux. It's said that Linus wanted a better Minix, but before that were the Berkeley Software Distribution and the GNU Toolset.

It's not at all clear why Linus is singled out for credit.

Re:Why credit Linus

[a_n_d_e_r_s]

Linux wanted have his own UNIX-like computor. Buying a VAX to run Berkeley Software Distribution was at that time not afforable for a mere student and you also had to have an AT&T license for them.
The 386 BSD was released after Linux was started; Linux was started in '91 and BSD 386 came out '92.

Also you had the large lawsuit regarding BSD in '92 which slowed the development for BSD versions for 2 years.
Since then BSD systems more or less has been playing catch up with the more capable Linux system.

Had the BSD for 386 been released earlier and has not the big lawsuit stopped the distribution of BSD for 2 years
that Linux would probably not been much more then a hobby project that become abandoned when something
better came along. But instead Linux become the #1 UNIX-like operating system of choice.

Re:Why credit Linus

[dirtyhippie]

Yeah, Linus didn't really do anything special here. The BSD folks were doing open contributions long before Linux did. BSD just wasn't available for 386 at the time. So Linus's contribution was making software that was able to bring that ethos available to more people while BSD was being ported and going through legal troubles.

The people who really deserve the credit are the BSD and ATT folks. Sorry.

Which of Linux or Git is the more important?

[Anonymous Coward]

Which of Linux or Git is the more important? Are trolls alive and well or what?

Git wasn't the first DVCS and should Git go away right now we would have Mercurial etc. to instantly fill the gap. We could switch Linux *tonight* to hg and have kernel developers be productive tomorrow.

What would we have to replace Linux in the millions of Android with overnight?

The guy who wrote that sentence is seriously high on dope: me wants some of what he's smoking.

Re:Which of Linux or Git is the more important?

[BitZtream]

What would we have to replace Linux in the millions of Android with overnight?

NetBSD if your primary goal is portability.
FreeBSD if your primary goal is speed, reliability and general overall niceness
OpenBSD if your primary concern is security or you just like self inflicted pain.

There is nothing unique about Linux or GIT, nor Linus fanboys ... they are just that, fanboys blinded by idol worship and unable to see clearly enough to realize both the things mentioned aren't unique.

Linux and Git are both just OSS copies of existing DCVS systems with nothing truely innovating about them.

Linux is a very impressive project, but thats not unique. Git really is nothing special and really shows the fanboy in someone. I've learned if your drooling over Git, 9.9 times out of 10, its cause you're a fanboy, not cause the benefits of Git offer something they couldn't get else where 10 years ago.

Re:Which of Linux or Git is the more important?

[serviscope_minor]

There is nothing unique about Linux or GIT, nor Linus fanboys ... they are just that,

You're talking utter piffle. Linux and git certainly do have unique features, just as pretty much all other operating systems and VCS systems have unique features. For instance, Linux is the most scalable OS available. It was (for a time) the only OS to support x86-64, too. And the first mainstream OS with userspace filesystems. Git is, or was when it was written, the fastest DVCS out there which seems useful for large projects. Not a feature I need, but that doesn't mean it isn't unique.

they are just that, fanboys blinded by idol worship and unable to see clearly enough to realize both the things mentioned aren't unique.

Where as you appear to be an anti-fanboy blinded by irrational hatred and rage directed towards those fanboys. That isn't any better.

Linux and Git are both just OSS copies of existing DCVS systems with nothing truely innovating about them.

I must say that Linux is the uniquest VCS in existence. I know of no other DVCS systems with filssystem laters, sockets, networking stacks, framebueffers, sound drivers, ...

Poor argument

[gr8_phk]

So he argues that Linus "invented" some development model that is somehow different than Stallman had in GNU. He even quotes from the original GNU announcement where RMS ask for code contributions. Yes, GNU was managed in a top-down way, where Linux *could* be claimed as bottom-up with Linus having the last word. That's the only distinction he seems to make. Openness and public participation were present in both, but because FSF was hiring people and paying them to do work doesn't mean they didn't have the same model. When you're a completely public project, you can reject contributions and turn them into wasted effort. When you're paying people to do work, you tell them what those decisions are going to be from the start so you don't waste money. I give Linus a lot of credit. Who wouldn't want to have his practical achievement under their belt? I give Stallman even more credit. Who wouldn't want to have his philosophical AND practical achievements under their belt? OK, that's a loaded question around here...

Re:Poor argument

[nomadic]

At least the OSI guys haven't shown up on this thread to accuse everyone of ignoring that they were the first to invent open source development in 1998.

Re:Poor argument

[Anonymous Coward]

At least the OSI guys haven't shown up on this thread to accuse everyone of ignoring that they were the first to invent open source development in 1998.

Come on! Give esr time to recover from his hangover!

Re:Poor argument

[Anonymous Coward]

Yes, GNU was managed in a top-down way, where Linux *could* be claimed as bottom-up with Linus having the last word. That's the only distinction he seems to make.

Small yes, but a critical one. If you think about it, all innovation in GNU (in the beginning at least) was done by RMS himself, namely, inventing the concept of GNU. The GNU tools were to be copycats of well known Unix tools. Nothing new here.
On the other hand, Linus was clearly calling for innovation and ideas.

Linus first ??

[Anonymous Coward]

Surely Linus was far from the first to promote "open innovation" ?!

CVS

[Per Abrahamsen]

I actually think CVS did more for "open innovation". Together with Sun sponsoring the various SunSites.

CVS was the first (at least widely used) free server based version control system, and it made it very easy for anyone with a server to setup a free software project. The SunSites were probably the most common hosting platform until SourceForge. Before CVS you either gave collaborators login access so they could work locally on your machine (GNU did that), or relied on sending patches, which Linus did for years. CVS made it so much more convenient. Especially with anonymous CVS which essentially allowed anyone to create their own "fork" that still tracked mainline. A very poor mans github.

CVS was buggy in design and replaced by SVN, and the DVCS's provided another leap ahead in collaboration, so CVS got a bad reputation. But for its time, it was a revolution at least as important as git.

Hell yes

[Anonymous Coward]

It certainly is! Just the other day my neighbour got me to help out her daughter who has a new camera and wants a photo-album, me being the go to guy for all things technical of course. At first she was asking about Picassa, the Windows photo thing and some other proprietary rubbish...after I had picked myself up off the floor from laughing, I suggested she installed a professional tool like Git to manage her photos for her. Her workflow now is really, really simple (she was also using Photoshop, I converted her to GIMP) - edit photo in GIMP, git add ., git commit -a -m "Changed contrast etc". It took an hour or two for it to sink in, but I think she got it at the end. When I left though she touched me on my shoulder, do you think she likes me? I'll take some flowers round next time and some cakes.

Linus?

[Exitar]

Charlie Brown's best friend?
I believed his gift to the World was The Great Pumpkin! (more than 50 years before the Flying Spaghetti Monster)

Three in one

[thaig]

Well, now we have the Holy Open Spirit and the Son so who's the daddy?

Re:Three in one

[Bill_the_Engineer]

Two cups of coffee and still no clue. Let me know if you figure it out. :P

Re:Three in one

[jd]

No idea, but the mother would be Inanna.

Re:Three in one

[Gubbeson]

Two cups of coffee

Well, there you go. Caffeine is what makes the brain tick, so it is obviously the father.

SHARE

[wandazulu]

Considering it's IBM's 100th birthday, it should be pointed out that a lot of the concepts TFA talks about were being done by groups like SHARE long before Linus was even born.

Re:open innovation = academic world

[Anonymous Coward]

"This moronic behaviour by Torvalds (and there is lots more) is one of the reasons I switched to FreeBSD."

Yes, thát makes sense!

No he didn't

[Anonymous Coward]

There are 6 billion people on the planet and many of the political and commercial processes you see came about as a result of the complex interaction between them. Picking out one person as the one true hero behind any project is stupid and childish.

People have discussed and argued for open government since forever. The Greeks had direct democracy more than two thousand years ago where citizens could vote for the legislation itself - the only kind of vote worth having!

I am glad Linus is around and he has contributed a lot but your analysis is nonsense.

And when you get that pipe out your mouth...

[Anonymous Coward]

Seriously, you're chowing down on Linus was seriously OTT.

The reason for Git was that Linus was wrong and RMS was right on the utility of a closed source program being used to coordinate development of the Linux kernel.

Torvalds would be the first one to say

[Anonymous Coward]

That all he did was start writing another OS and then continuing to support Linux with the help of thousands of other programmers.

Some of the techniques he used might have been innovative, but the broader innovative ideas had already been put into place by RMS and Tanenbaum.

Without Minix and without GNU compilers that had already been written to jump start the writing of the Linux kernel it would have been really difficult to write the first Linux kernel. Without the GPL Linux would have never been successful, it would have just been another failing bsd unix without the GPL.

redefining the meaning of pioneer

[doperative]

`Linus Torvalds has done well with Linux and now Git, but I don't really seeing him deserving the title of pioneer of "open innovation". At least not in the way the author is using the word "pioneer"'

That might be true but first one would have to redefine the meaning of 'pioneer', 'creating', 'taking advantage of', "open innovation" and "open source methodologies"

"I'm doing a (free) operating system .. and I'd like to know what features most people would want. Any suggestions are welcome, but I won't promise I'll implement them :-) link

Open Innovation as an emergent property of FLOSS..

[borgheron]

All Linus did was write a kernel and all of the things that the article credits him with inventing, were already part of the free software landscape prior to his posting to the minix group.

How do I know? BECAUSE I WAS THERE. I remember the posting on the minix group and I remember the first versions of Linux being passed around University of Maryland when I was going there. This so called "Open Innovation" is an emergent property of Free Software. So, please, get your facts straight, and stop your hero worship.

GC

More corporate

[Jonathan]

While I prefer git, it's worth remembering that just a couple of years ago, svn was the young rebel upstart (hence the name "subversion") -- most commercial shops were using cvs until recently. It's a hard sell to get IT departments to shift again ("why, just a few years ago you nerds made us switch to svn when cvs was working just fine. Now you want to shift again?")

Re:More corporate

[asdf7890]

SVN is still widely considered one of the best at what it does, so if it works for your current requirements there is probably no need to switch. A lot of companies using it will remain there because they have no particular reason to move (i.e. the shortcomings that can be resolved with something else like git don't affect them in a way they are aware of) and it is integrated into their key processes like build and test.

Companies moving from something else (something proprietary and/or order than subversion) or starting from scratch are more likely to go for git though, which I think is where most of its current momentum comes from.

The interface(s) to git are a major factor that will keep some companies away for some time. While we experienced unix-a-like people are perfectly fine with it and other with a decent brain between there ears will pick it up quick enough too, there are a lot of people out there who won't cope with it as-is so many companies will keep back until there is a nice well-rounded stable official GUI (preferably that integrates well with VS in the Windows world). And it isn't just junior/bad developers to worry about - in a small company the PHB will expect to be able to look at what you are working on so will want access (in a larger company there are several levels of shielding between the PHBs and the source repositories, but not in smaller ones) and I don't know about you but I don't fancy being the one who has to try explain git to my line-manager's boss...

FYI: I'm starting to use git for my own stuff, but I'll not be recommending it for here just yet.

Re:More corporate

[cynyr]

I also would think that git-svn would be used a lot by the devs to talk to the central SVN repo.

Re:More corporate

[jbolden]

Subversion was a minor upgrade from CVS. Essentially it should have felt like a major version upgrade. GIT is an entirely different philosophy of source code management. Its a fundamental shift, and one that for most commercial shops really doesn't make sense.

Re:More corporate

[asdf7890]

That is something on my list of things to investigate, once I've kicked git around a little more and feel confident I'm understanding the semantics of all its basics.

Re:More corporate

[thePuck77]

Subversion was a minor upgrade from CVS. Essentially it should have felt like a major version upgrade. GIT is an entirely different philosophy of source code management. Its a fundamental shift, and one that for most commercial shops really doesn't make sense.

Why do you feel Git doesn't work well for commercial shops? With a lot of modern coding practices (Agile,etc), I feel like it's the only one that makes sense.

Re:More corporate

[jbolden]

The core idea of distributed source code control is distribution, that there isn't a central repository and a central administration and management structure explicitly handling branches and merges. In most commercial shops you want non distributed, you want a central administration and management structure.

Ideas like file locking are a plus:
Joe has these files, Samantha is working on this module, and Michael on that one. Joe tries to change Michael's file bang runs into a lock; that's considered a good thing.

Mercurial = Git in that regard

[Anonymous Coward]

While git did that, mercurial which was developed at the exact same time does exactly the same thing. Both learned from the experience made in managing the Linux kernel with Bitkeeper where most of the ideas concerning distributed version control originated, of course. The time was ripe.

git real

[Anonymous Coward]

Git is the RCS of distributed version control. One day, someone will write the Subversion of distrubuted version control. With an intuitive command line interface and sane default behaviors; I can't wait.

But even that won't be anywhere near the importance of Linux.

Monotone

[jonabbey]

Linus imitated Monotone more than he imitated BitKeeper. BitKeeper is a DVCS, alright, but it is based on an SCCS (!) underpinning. Git is based on a content-addressable hash model, which Monotone introduced into the version control space way back when.

Linus' = plural possessive of Linu.

[Anonymous Coward]

The correct singular possessive is Linus's (http://www.bartleby.com/141/strunk.html#1). The s-followed-by-apostrophe for a singular plural is reserved for Jesus' . Don't confuse the two.

Re:Linus' = plural possessive of Linu.

[polymeris]

...a singular plural...

Wat?

Glyn's usage is correct, no matter what the Chicago manual says.

"Open Innovation" - rediscovery

[Zoxed]

I think if you take a longer historical perspective then "Open Innovation" is what humans have normally have engaged in. It is only quite recently that patents and copyright etc. came on the scene to create "closed" environments.
Imagine what would have happened if someone had patented fire, metallurgy, selective breeding (plants and animals) or copyrighted the musical scales !. Instead ideas travelled around openly, anyone could experiment, copy ideas etc.

Re:"Open Innovation" - rediscovery

[Anonymous Coward]

Indeed its hardly new: See Joel Mokyr's book "The Gifts of Athena: Historical Origins of the Knowledge Economy", Princeton UP. 2002. Or his paper “The Intellectual Origins of Modern Economic Growth,” Journal of EconomicHistory Vol. 65, No. 2 (June 2005), pp. 285-351.

He just do things like Swedes/Finns usually do

[Anonymous Coward]

Linus Thorvalds seem to be very good at what he does, but he haven't invented anything new, he just do things in a manner that seem new to people outside Sweden, Norway and Finland.

[Linus Torvalds is an ethnic Swede, from Finland. Swedes are the second largest ethnicity of Finland.]

Re:Open Innovation as an emergent property of FLOS

[Urkki]

All Linus did was write a kernel and all of the things that the article credits him with inventing, were already part of the free software landscape prior to his posting to the minix group.

How do I know? BECAUSE I WAS THERE. I remember the posting on the minix group and I remember the first versions of Linux being passed around University of Maryland when I was going there. This so called "Open Innovation" is an emergent property of Free Software. So, please, get your facts straight, and stop your hero worship.

It's emergent same way as democracy is emergent in human society. It's bound to happen, yet it's not automatic, and often it takes blood, sweat and tears to make it happen. Anybody who can pull it off and become "hero" needs luck of course, but luck alone will not make it happen. Minix didn't take over the world, it was left in the dust of history while Linux made history.

Linus should receive more notoriety.

[DrStoooopid]

I think he should be named man of the 20th century. How one man would go against the grain, write his own OS to thumb his nose at "The Man", and just look at how much he's contributed to the world without even trying or expecting anything in return. He's like Gandhi of the digital era.

people always have to turn it into Linus vs. RMS

[zhub]

Maybe everyone will be happy when we just start calling it GNU/FOSS.