There's an interesting point hiding in the virulence of this post:
Canonical engineers work on Ubuntu.
Red Hat and SUSE have historically hired swaths of "upstream" developers, often just to work on their own projects and propel Linux along. When I visited an Ubuntu Developer Sprint a few years ago, as a former KDE developer, I was surprised at how few people I knew. It wasn't that they weren't engineers; it was simply that they were working on Canonical projects: Ubuntu, Launchpad, Rosetta, etc., not j-random-OSS-project.
The interesting thing is that they've been trying to hire more upstream folks. Most core OSS developers don't have a problem finding a job. They've also been at this a relatively short period relative to Red Hat (and SUSE), so there's would be no surprise, even if the hiring practices were similar, that the total commit count would be far smaller.
I can never understand some elements of the Linux community to speak ill of other community members success.
Ubuntu is probably the best thing thats ever happened to the Linux desktop - its being installed by manufacturers the world over and thats years of contributions that are actually being used by people.
I'm at a bit of a loss to what Greg feels Canonical should do about this. Stop offering Ubuntu? Have Mark post a warm and fuzzy blog post? High-five Redhat engineers every time their code works?
Canonical have done a monumental service to the open source community via their Launchpad service. They provide space for documentation, VCS hosting, issue tracking, repository hosting and build services. They enable communities.
I'd also like to make a special mention to the thousands of hours Canonical engineers and Ubuntu community members spend pulling patches from "downstream" (other distributions including Ubuntu) and pushing them upstream (the kernel, GNOME, KDE, Xorg, etc) - this sort of behaviour fosters better software for the whole community - not just Ubuntu.
I know of people who do push patches back to upstream. (I used to spend a lot of time doing it.) But, last I checked, this wasn't really standard operating procedure.
Am I wrong? I sure hope so.
But I can't help but notice the Utnubu [1] project still exists. And still isn't Canonical supported.
"If you doubt, for a nanosecond, that Canonical is a marketing organization masquerading as an engineering organization, then you’re either an unapologetic Ubuntu fanboy or you’re not paying attention."
1. Poisoning the well is an ugly debating trick.
2. He's somehow implying that marketing FOSS is not a useful contribution.
As for the numbers:
1. "$SUBJECT is the percentage of contribution to the GNOME codebase." ... actually, no, it's the percentage of commits. I'm not refuting the value of that metric, but misrepresenting what's measured is again a debating trick.
2. Red Hat has about 10 times the number of employees that Canonical has. Taking economies of scale into account, is the 16:1 ratio so bad?
Finally, given how well Red Hat's doing, should they really get into a fuss over any supposed free-rider phenomena?
> He's somehow implying that marketing FOSS is not a useful contribution.
Whether or not it's useful, the FOSS community does have some holy ground that shall not be trod upon. The extent of that holy ground varies, but one of the basic tenants of open source is Credit Where Credit Is Due. Generally speaking, the community is very flexible, but don't be surprised if this is where the gloves come off.
NO mention of "linux" or "GNU" anywhere on the page. Click on the "How Can It Be Free?" link, and no mention of GNU or the GPL.
In fact the reasons they give are "1. It's Open Source" and "2. It's Managed and Funded by Canonical". Open Source is a side-effect of Free as in Freedom, a concept that Canonical is agnostic about. And I'm not sure what point 2 means.
I can understand not wanting to turn people off with scary-sounding words, but even Mac like to tout the "rock solid unix foundation" on which OSX is built.
I think they view GNU and GPL explanations as not very friendly to their perceived audience which is the common computer user. Branding it Ubuntu in their view makes it simpler to understand. I think they want to be like OS X and relegate discussions of BSD and Mach Kernel to some really hard to find page.
At least that's how i'm interpreting their pr stuff.
In all respect, then perhaps the FOSS community should change perspective. Ubuntu has been one of the best PR-moves that Linux ever had. If the goal is to get people to use the software, then they contributed a great deal to the project.
The problem with that semi-assertion is that the goal is differs greatly amongst OSS developers, from "I did something useful for me and am sharing it in case it's useful to anyone else," to "I'm going to take over the world by getting OSS into everyone's hands." ... and everywhere in between. Canonical clearly tends toward the latter end of the spectrum, where I imagine a company like Red Hat is somewhere in the middle.
Thanks, that's a good point (I should read Homesteading the Noosphere again :)). I guess I've not been aware of (or sensitive enough to pick up) any clear offences by Canonical in this respect. Maybe that's because I only subscribe to their security notices, not the general news :)
We use CentOS all over the place in our development and internal infrastructure. But when we have to feel confident that a server can be maintained and secured reliably we happily pay our RHEL fees.
I think that RHN is a great value add. Letting me updates servers from a webpage at home. If you aren't a small company and have the budget, RedHat is worth the money in my view. But for non production, CentOS would be my choice too.
It is not free-riding. Canonical cares a lot more about user-experience and the little things. It classical capitalism: everybody does what he thinks is best for her and the end everbody profits.
Red Hat abandoned the desktop community/didn't take it seriously and then are surprised that someone who does gets all the mindshare and community goodwill?
It seems to me to be a miscalculation to say that the desktop hobby market has no money in it. All those younger school and university students which Ubuntu fosters will eventually be in positions where they get to choose what distribution is installed on their servers at work and it will be Ubuntu or Debian, and when GNU/Linux is rolled out on corporate or educational desktops these people are going to choose Ubuntu most likely.
I'd say either take the desktop seriously or re-evaluate your spending on it. Don't do a half-hearted job and expect people to love you. I understand this is probably the engineers getting frustrated at decisions which have been taken by the managers at RH.
I remember when RedHat dropped their desktop version (ah, Redhat 9...). At the time I remember thinking is was a dumb move. That having a good, well supported enterprise desktop was too important to RedHat and Linux in general. Then they did their half-hearted spinoff into what is now Fedora, and for a while it seemed like it desktop Linux languished.
Now, I'm pretty sure it was the right move. Enterprise Linux pays the bills. There are very few major installations of Linux desktops, and that's about the only place where you could extract revenue.
I think that they decided that they couldn't make both the enterprise and desktop users happy. Enterprise servers want to be stable and predictable. Desktop users want the new shiny. They had to choose one. So they went out, bought a suit, and went with the ones that gave them more money. Based upon their current financials, it's hard to argue with that strategy. And the funny thing is, they still managed to contribute more code to Gnome than Ubuntu.
As a side note: I used to run all Red Hat servers... then dabbled in Gentoo but recently I've used all Ubuntu. Just this past week I've run into some issues with some very important software (to me), that seems to be related to something in Ubuntu. It runs perfectly in the more conservative RHEL 5, but fails on the more recent Ubuntu. This has made me think a lot about what I'll be using for my servers in the future...
You walked almost the same path as me then (re. distros), so I'll give you an advice: Debian. Ubuntu was never a proper server distribution. With their lack of own developers, majority of serious problems have to be pushed upstream as Debian issues anyways. If you liked the "conservative" RHEL, just go for Debian stable. It's the best of both worlds and it is what I'm happy to recommend these days.
The problem with Debian in the past (and I say this as someone who formerly maintained some packages and is a Big Fan) was that the release schedule was unpredictable. I don't know what that's like these days - anyone care to comment?
Seconded, though I'm more interested in Canonical's commitment to Ubuntu LTS releases: at work I don't want to have to upgrade distributions every two years (or six months: Fedora, Ubuntu non-LTS). The same goes for RHEL/CentOS. These distributions are attractive to me in a small business because of the support commitment of the organizations behind them.
The point is (and it's actually written in bold at the middle of the article) that Red Hat invests a lot more in the desktop than Canonical, while Canonical reaps all the benefits. Your argument that Red Hat is not treating the desktop seriously seems a bit shallow when comparing the amount invested by each company in desktop development.
Yeah, but Ubuntu's PR has always promoted the desktop (and laptop, and netbook..) They now talk about the server more than they used to, but ubuntu.com still gives the desktop a lot of emphasis.
Meanwhile:
* redhat.com's frontpage is devoid of anything but the odd "Desktop" link
* Way back in 2003, Red Hat's then-CEO said consumers should buy Windows [1]
* In 2008, Red Hat stated that "we have no plans to create a traditional desktop product for the consumer market in the foreseeable future." [2]
* In 2009, their CEO said "the concept of a desktop is kind of ridiculous in this day and age." [3]
I think you could say that Red Hat's focus makes sense, or even that their view on the Linux desktop is more honest. But Canonical, at least, has aspirations for the Linux desktop, while Red Hat decided a long time ago that they're not really fussed.
Credit where credit is due is one thing, and in [2] Red Hat lays claim to the work they put into the desktop, and its suitability for certain niches. But contrasting their PR with Canonical's, can anyone be surprised the "power user" Linux desktop set don't give Red Hat much attention?
And still... That's not the point. The point is that Canonical is targeting the desktop market, but their contributions seldom get upstreamed. OTOH Red Hat makes most of the contributions and they're not even focusing on that market.
Redhat is a much larger company with much larger resources. The contributions they are able to make is thus much
larger. I don't see how this should be taken as a surprised.
Their contribution to the desktop are nothing compared to their server side work.
Most of what Ubuntu/Canonical contributes is integration work which is specific to their distro in nature. That's already a lot of work for the size of their company. If Redhat wanted to they could promote Fedora as a desktop system. I know many organization that use it but for the man in the street (MITS) Ubuntu is just more visible. Red Hat could put some PR money on Fedora.
I see also a lot of work being put by Canonical in just replicating work that took Red Hat years to make. Building their enterprise infrastructure is sapping a lot of work. Landscape,Launchpad,Ubuntu One... (Launchpad could be dropped considering the other services popping up that are much better.) are expensive projects to make and aren't of much use to the FOSS community at large.
Red Hat concentrated on the enterprise. They certainly haven't abandoned it - they bought Qumranet as much for their desktop virtualization solution as the KVM technology.
These numbers show that Red Hat did the right thing too - they're a highly profitable company, and they're bringing massive value to the community. I really don't understand why you seem so aggrieved at a company which contributes so much.
They are highly profitable? Thank god. I don't know enough about them as a company, but it is really good to know they aren't teetering on the brink of destruction. They are currently in my list of companies I really want to stick around.
I had some investments in the company for a while. It seems to be very well managed. There revenues have had a decent growth over the years I've followed it.
eventually be in positions where they get to choose what distribution is installed on their servers at work and it will be Ubuntu or Debian
Eventually is a very long time. I ought to be there by now, I've been using Linux and Unix in general professionally for 15 years now. But the Linux I use in production is dictated absolutely by what various vendors certify their products on. That's true for anyone doing anything more complicated than "a website".
I think you're dead right. I use Ubuntu on the desktop, and I would specify it for any server installation. Not because I dislike Red Hat, I'd just rather use something I'm familiar with.
> Red Hat abandoned the desktop community/didn't take it seriously and then are surprised that someone who does gets all the mindshare and community goodwill?
I don't know how a statement so riddled with misinformation got upvoted so high. Your comment is not only wrong, it goes out of it's way to spread FUD. So much wrong, I just don't know where to begin.
Oh, I'm sorry. I was unaware the commenter was referring to a post made by Red Hat rather than someone not from Red Hat. I wasn't aware that Red Hat no longer contributed to desktop linux.
"If you doubt, for a nanosecond, that Canonical is a marketing organization masquerading as an engineering organization, then you’re either an unapologetic Ubuntu fanboy or you’re not paying attention."
I didn't know Canonical claimed otherwise. All the way back to Launchpad Bug #1: "Microsoft has a majority marketshare"[0], they've been about marketing and sales. Most of the software I've seen coming out of Canonical has been of the "hundred papercuts" variety[1] - little stuff that's easy to fix, but hard to get developer attention for.
I won't know about Canonical's claims but it is sure that they're perceived to have more impact on desktop than Red Hat, and that's a very fair perception regardless of the number of commits Redhat does on one particular piece of the desktop.
Linux has been starving of a good, stable desktop, and Ubuntu has provided that. Regarding free-ride, sorry but that's how open source software works. If Ubuntu credits itself for code that RedHat wrote that's not right, but if Ubuntu claims they changed the Linux desktop they have a very reasonable point regardless of who wrote what.
Isn't the point of open source that people can fork and do what they like with it? Canonical has made a major impact on Linux as being the gateway drug for so many of us. I can honestly say I would not be using linux if not for ubuntu.
Linux needs marketers, otherwise we just have cool software that no-one uses.
I learned several years ago that Redhat is probably the single biggest entity I have to thank for Linux, besides Linus himself. Start using RHEL, and as you begin to deal with bugs and such, you notice just how much is going on, and just how much goes upstream.
This has not changed, and generally speaking I hope it won't change; I don't spend much time browsing raw code, but they seem to produce the highest quality results.
This is a little off-topic but to me there's no significant differences among the main general-purpose Linux distros.
I help manage a ton of a mixed RHAT, Centos, Debian and Ubuntu servers and the day-to-day operations are the same; we just install apt in the RHAT-based servers since it's nicer (sure, there's the chkconfig vs rc-update and /var/log/secure vs auth or mail.log vs maillog, big deal).
I'd install or recommend one or the other based mostly on the versions of the applications (apache, postgres etc) that are packaged, if anything. Otherwise I say go with what you are familiar with or what your friends/staff are familiar with.
That's on the server side. On the desktop distros (which I don't use much) it's almost the same. I spent a few years moving from distro to distro and I came to the conclusion that the difference between a new and a previous release of the same distro can be bigger (for me) that the difference between two current distro versions, so again it doesn't matter or rather any little thing can tip the balance to decide one distro or another. It comes to a point where I install say Ubuntu in my laptop and WiFi doesn't work and I just install RHAT or whatever; it's faster to do a re-install with another distro than to troubleshoot the problem.
> If you doubt, for a nanosecond, that Canonical is a marketing organization masquerading as an engineering organization, then you’re either an unapologetic Ubuntu fanboy or you’re not paying attention.
Call it what you will Ubuntu's success has everything to do with the default install being quite usable by end users. So what should we call that? Neither marketing engineering sounds right.
If you want people to use your software you have to invest in good engineering and you have to invest in good marketing. They're both crucial. Engineering alone doesn't cut it, nor does marketing.
Canonical's work is just as important, if not more so, in putting GNOME in the hands of users.
The business model of Canonical is very simple and clean - themed, non-advanced-user-friendly version of debian (reuse) along with active promotions (money).
Why they should write code if their main activity is decoration and customizing?
Fedora people are the ones on the bleeding edge. RedHat is the second recycling of their effort, CentOS - the third. ^_^
Debian (especially) and Ubuntu are famous for their conservative, slow-moving, less-effort approach.
Microsoft has hugely more code in the Linux Kernel than Canonical.
After MS got their 22Kloc of Virtual Machine related driver code up-streamed, they massively outweigh Canonical's meagre to non-existent kernel contributions. Canonical just don't do infrastructure work, more or less.
It's also true that Canonical aren't very good at persuading their various up-streams of the value of accepting their patches. It's also true that they don't really do much general linux/gnome infrastructure stuff - they mostly concentrate on Ubuntu specific UI/UX stuff - which is very rarely adopted by upstream projects, for various reasons. It's also true that litl (http://litl.com/) a smallish, 40 person, single product start-up, has about the same level of Gnome contributions as Canonical, a company ~10x their size: (http://blogs.gnome.org/bolsh/2010/07/28/gnome-census/)
I'm not sure of the wider meaning of these things is, really, if they have any; especially considering that I've currently got Ubuntu installed on all my computers.
Canonical engineers work on Ubuntu.
Red Hat and SUSE have historically hired swaths of "upstream" developers, often just to work on their own projects and propel Linux along. When I visited an Ubuntu Developer Sprint a few years ago, as a former KDE developer, I was surprised at how few people I knew. It wasn't that they weren't engineers; it was simply that they were working on Canonical projects: Ubuntu, Launchpad, Rosetta, etc., not j-random-OSS-project.
The interesting thing is that they've been trying to hire more upstream folks. Most core OSS developers don't have a problem finding a job. They've also been at this a relatively short period relative to Red Hat (and SUSE), so there's would be no surprise, even if the hiring practices were similar, that the total commit count would be far smaller.