APLawrence.com -  Resources for Unix and Linux Systems, Bloggers and the self-employed

Succession Matters for Linux (and everyone else)

At Linux in the long run I expressed the opinion that Linux in general could suffer when Linus Torvalds steps aside or dies. That opinion is generally unpopular with the Linux community, but I think it's defensible.

Apple

But before getting to that, I need to explain that I do think that Linux has a lot less to fear from succession than Apple does. Right now, Apple is riding high, gaining ground regularly and stealing both Microsoft and Unix users right and left. It's not just iPods and iPhones; OS X is a power to be reckoned with. I can guarantee that none of this would have happened without the vision and strong control of Steve Jobs. And when Steve is gone, my bet is that Apple will slide back to near oblivion very quickly.

Oh, there will be momentum. Jobs has built a great product line and that could pull them along for a few years without him. But unless someone equally strong minded and equally gifted takes the reins, Apple will falter.

Microsoft

I think Microsoft is going to suffer no matter what. What they actually need is a Steve Jobs, but their board isn't likely to appoint one even if one could be found. However, "suffering" in Microsoft's case might mean slipping a bit: if Apple and Linux fall behind, and nothing else fills the void, Microsoft junk can continue to reign just because nothing else is there to fight them. Let's hope that is not the case: if Apple and Linux can keep up the pressure, Microsoft WILL slide downward.

Linux

The "succession doesn't matter" arguments boil down to demonstrating that Linus Torvalds doesn't do all that much now, so succession really won't change much. Defenders are also quick to argue that Linux is only responsible for the kernel anyway and that most of the obstacles to widespread popularity come in other areas, like the Desktop. Those arguments definitely have truth in them, but there's more to it.

First, the "it's only the kernel", because that's the weakest: without excellence in the kernel, everything else falls apart. You can't ever hope to have the slickest, most user friendly, absolutely the grinchiest desktop if you don't have a strong and reliable kernel underneath. If the Linux kernel gets screwed up by politicking, in-fighting and neglect, Linux as a whole will suffer.

But it won't, right? It's all decentralized now, it doesn't really need Linus.

Maybe. I want to believe that as much as anyone else, and perhaps even more so: I think open source is critical to our future and that we will lose a lot if proprietary systems regain their absolute dominance. But I think that the "nothing to worry about" arguments neglect two basic elements.

The first is charisma. Linus Torvalds IS the face of Linux just as Steve Jobs is the face of Apple. Outside of the kernel mailing lists, nobody knows any other name. That's important, because perception is reality: if the business world sees Linux as leaderless and rudderless, faith will be lost. Again, here it's not a matter of reality, it's perception that matters. Even if a strong leader steps up to take the reins, he or she will be an unknown, and that can shake confidence.

Secondly, I suspect it's unlikely that a strong leader will emerge. There are too many strong willed individuals capable of taking over, so there likely will be extreme bickering and fighting. That may resolve without serious damage, but it could also lead to dreaded forks or worse, management by committee. And that last is probably the thing that should be feared the most. A committee won't screw up the kernel by making bad technical decisions as a popular but incompetent leader might; no, they'll screw it up by inaction.

That's the real danger here. A moribund kernel lagging behind other developers needs will either push them to other platforms or tempt them to fork.

None of this HAS to happen. My hope is that it WON'T happen. But it CAN.



Got something to add? Send me email.





(OLDER)    <- More Stuff -> (NEWER)    (NEWEST)   

Printer Friendly Version

-> -> Why Succession does matter


14 comments



Increase ad revenue 50-250% with Ezoic


More Articles by

Find me on Google+

© Anthony Lawrence







Sun Apr 6 13:50:20 2008: 3971   Prash


Hello again! :)

How many people have actually heard of Linus? Only those interested in Linux as an idea as opposed to just an OS. It seems to me that a major reason for his fame is that the project is coincedentally named after him. That's not to belittle his work or to suggest he is egotistical. He seems humble enough not to have called it Linux if he knew it was going to take off from being a hobby into a major OS. I can't name the "head" of any other similarly successful open source project. Partly because that person usually isn't much more important than anyone in the top 5-10. And secondly because they don't all need to define a single person as leader as the role can be somewhat meaningless in a collaborative process (e.g. Apache Software Foundation). As I understand it there are 2 main candidates lined up as potential successors. Andrew Morton and Alan Cox. For the people to whom it matters these names are already as well known as Linus' and there won't be a messy power struggle between them as their power is defined by how much the community respects them.

The reason I'm arguing this so vigorously is that the stability of open source over proprietary development is one of its major assets and should be one of the reasons that you, as a self proclaimed Linux pusher, should be pushing it. Of course ANYTHING can happen. But it's only meaningful to talk about what is likely to happen based on an understanding of the situation.

And by the way, I think you overestimate the role of the kernel in desktop slickness. The Linux kernel is miles ahead of Windows kernel yet Windows is more slick. Sure the desktop environment is useless without the kernel but on a home computer the opposite is just as true. The future of Linux on the desktop now lies in the hands of those that work on the desktop environments, sound servers, device drivers, browsers, media players and the distributions that put all these nuts and bolts together in an accessible way.



Sun Apr 6 17:24:25 2008: 3979   TonyLawrence

gravatar
Actually, I think a lot more people know Linus Torvald's name than have ever heard of Cox.

Apache is, I think, and Firefox, are examples of how even really bad applications can still be popular for other reasons (I run this website on Apache and will not consider anything else in spite of the fact that its programming interface is awful; I use the Fox every day while cursing it for its instability).

But you might be right that I am overestimating. It's one of those cases where I will be very happy to be wrong.



Sun Apr 6 17:42:43 2008: 3980   TonyLawrence

gravatar
Also remember this: it's not just THIS succession that matters, but every one after. If Morton or Cox take over successfully, great, but they are mortal also - the crown must pass again.







Sun Apr 6 17:59:46 2008: 3981   TonyLawrence

gravatar
An older post that touched on some other problems that can come from succession: (link)



Mon Apr 7 23:32:53 2008: 3990   drag


What I think is more likely, hopefully more likely, is that people will eventually stop caring.

'Open Source' and all that existed long before Linux came along. It just wasn't as formal, I suppose.

So I think that Linux's great contribution to computer science, or maybe to society in general, is going to be making the OS largely irrelevant to people's day to day lives. Even that portion of people's lives that involve computers. It'll end up being unknown by the vast majority of people that will use it or program on it. This will lead to it's ultimate destruction.

Sure the pace of development is fast and furious right now... but 10 years from now? 20? When you have a kernel that is flexible enough to work on any sort of new hardware platform people can think of, and make it mostly trivial to develop drivers for, and it'll efficiently scale from wristwatch computers to a super computer with 10s of thousands of cpus with the flip of a couple compile-time switches... who really gives a crap anymore? It's good enough and people will maintain it because there will be a financial need to, but it won't be anything special. Everybody will just use it because it's cheap and it's there and it works.

When it gets to the point were it's about exciting as going down to the hardware store to buy a 4-inch long bolt then people will use it a platform to doing more interesting stuff.

Like... I donno... Say having a entire corporate network that is ran like a single OS. Having separate OSes for most devices will seem as artificial and compartmentalized as a Unix machine with C: and D: drives. With a single network-wide kernel and with devices that don't even need to boot or load a OS to run. Personal computing devices that simply 'turn on' when connected and share their resources of the entire network... sort of like a mouse does when you plug it into your USB hub. CPU time and memory is shared. No storage is even needed, it just exists in the memory of the devices connected to it. A network that is self repairing and self configuring that will require very very little administration. It'll grow and shrink to met demands of it's users automatically. Oh.. and with good transparency and strong security.

That's what I see as the long term fate of Linux. Not that it's going to fall a apart due to organizational failure.. but that it just ended up getting _boring_. Maybe even stagnate and disappear, but it'll be largely irrelevant because a team of skilled programs could probably generate a effective replacement using _real_ computer-aided software development.

It's something that is going to be more and more common...
A modern example of this approach is Samba 4, much of it's code is automatically generated. I mean I don't understand much of it, but it sounds fantastic. Use IDL files to describe protocols and interfaces and use perl to generate the C code you need..
(link)
(link) <--- perl code



Mon Apr 7 23:41:18 2008: 3991   TonyLawrence

gravatar
Boring - I like it :-)



Tue Apr 8 19:40:28 2008: 4000   ShaneKerns


Tony, I really don't know what to tell a fool like you. If you find Firefox unstable then tell me do you find IE6 or IE7 or Safari more stable and might I add standards compliant?
I agree with you about the apache interface but its just a web server that can be enhanced when there is a real need for it. Besides its a great functional web server.

There are more than enough kernel developers out there perhaps even stronger than Linus that could take his place, they are just relatively unknown at this point. Besides all Linus does these days is review kernel code submitted by Google, Intel, AMD and several others and just essentially package it. Rest assured if Linus does pass away kernel development with still continue in much the same way as it does today.



Tue Apr 8 19:46:44 2008: 4001   TonyLawrence

gravatar
Shane - I'm not a fool. Firefox is far less stable than other browsers, but I don't care - mostly because of that wonderful "Restore Last Session" they added a while back. Let it crash - who cares?

As to Apache, you missed my point: obviously it is the ONLY real choice (though ROR is coming on strong). But remember its namesake "A Patchy Web server". It's an architectural mess and full of confusing crap.

And yes, I'd like to "rest assured". But I don't think it's quite as simple as so many of you insist.



Tue Apr 8 20:04:39 2008: 4002   TonyLawrence

gravatar
I'm just going to add one more thing here:

I think a lot of techy Linux folks don't see this because they ARE tech types. Tech people don't understand the social side of things, are often caught unaware by political issues that are so obvious to others.

Your succession is no problem arguments are technical. You don't grok the real world.



Tue Apr 8 20:16:43 2008: 4003   Jorge


I'd like to point out a few thingsi think you are missing in your arguments.

Argument 1: Linux lacking a face presents a problem

I think you are making more of a big deal out of this than is warrented. Sure a big face helps, but its not that important. Ask the basic user who runs Microsoft (and even some IT people) and see how many of them get that answer right.

Argument 2: It is unlikely that a strong leader will emerge after Linus
This is the argument that I find shallow. It sounds like pure speculation and guesswork. Do you know anyone who would be in the running to take over when Linus leaves, or are you just basing it on the broad stereotype of kernel developers?



Tue Apr 8 20:23:24 2008: 4004   TonyLawrence

gravatar
I'm basing it on the reality of the world.

Actually, I don't know why I even brought this up: Linus is much younger than I; by the time he hangs up his spurs I'll be babbling or dead (yeah, I know: you think I'm babbling now).

I liken this to the patent threats and the inviolate GPL: all you tech guys think Linux has nothing to fear from patents and that the GPL can never be challenged.

The real world runs on money and politics, not on logic.



Tue Apr 8 20:25:52 2008: 4005   TonyLawrence

gravatar
I'll let a few of the "no problem" folk have the last words and then I'll be closing comments here.





Wed Apr 9 02:39:37 2008: 4010   CalebCushing


There are other kernels which most of the existing FOSS software can/would/does run on.
Their are the BSD kernels, and the GNU HURD. As of right now I think linux is the best. But should Linux ever fall apart we are not without, alternatives or hope.



Wed Apr 9 10:16:58 2008: 4015   TonyLawrence

gravatar
That's true - we're never without hope - well, not unless Microsoft bribes every government in the world and gets them to outlaw open source.. or manages to patent so much that there's no wiggle room..

------------------------
Kerio Samepage


Have you tried Searching this site?

Unix/Linux/Mac OS X support by phone, email or on-site: Support Rates

This is a Unix/Linux resource website. It contains technical articles about Unix, Linux and general computing related subjects, opinion, news, help files, how-to's, tutorials and more.

Contact us





The idea of "work, then get paid" has been deeply ingrained in our culture by employers who want to limit their risk. Well, I like to limit my risks also. I like to get paid before I do work. (Tony Lawrence)

A Perl script is "correct" if it gets the job done before your boss fires you. (Larry Wall)












This post tagged: