Greg is, of course, a well respected contributor to the Linux kernel, having sustained a significant level of contribution over a period of several years. I’m grateful to him for his technical contributions, which of course benefit Ubuntu as a consumer of the Linux kernel. However, his contribution to the public dialog about the Linux ecosystem leaves much to be desired.
We all have bias, and the best that we can do is to disclose it so that others can take it into account when hearing our ideas. Unlike the presentations given by other Novell employees at this and other conferences, Greg’s slides omitted the Novell logo.
--Andre Gide
Comments
AlexH
2008-09-18 14:27:48
Nobody's disputing his figures, so I think it stands up as fair criticism, personally.
AlexH
2008-09-18 14:28:56
Anonymous
2008-09-18 15:05:28
Did you consider that he did it because he was presenting his own strong opinion rather than one of his company?
Baby In The Bath Water
2008-09-18 15:21:26
Note: this comment was posted from Novell's headquarters.
Baby In The Bath Water
2008-09-18 15:29:15
That's a pretty appropriate quote for this site, it applies very well to Roy.
Note: this comment was posted from Novell's headquarters.
bboissin
2008-09-18 16:09:05
BW
2008-09-18 16:39:26
What I find highly suspect is Greg's use of the "enterprise developer" label as being a bad thing. The latest Ubuntu release (2.6.24) uses a kernel a lot newer than either Novell (2.6.16!) or Red Hat (2.6.18) does in their enterprise release, but I don't see him criticizing either of them for dragging their feet.
Further, any genuine enterprise developer knows that it is much, much more difficult to get a backported patch accepted by enterprise distros like Novell and Red Hat if that patch is not first vetted by the LKML and accepted into the upstream. For instance, in the case of a patch that I developed (which you can find at http://lkml.org/lkml/2007/1/18/137), it took me well over 6 months to hunt down the bug, another two weeks to get into the mainstream - but then Red Hat sat on the patch for *over a year* before finally accepting it into RHEL 5.
So, IMHO quite reasonable to assume that anyone doing kernel development on Ubuntu will have to go through similar hoops to get a backported patch accepted there - so the upstream gets its patch, which might not have been written if someone wasn't using Ubuntu - but the patch isn't contributed under Canonical's name.
The only real meat to Greg's argument is that Canonical employees aren't contributing as much to the kernel, but quite frankly, who cares? Maybe they're better suited to doing GUI development, or some other area of Linux *as a whole,* including the desktop. The kernel is just one piece of the whole puzzle and to discredit an open source developer because the projects they contribute to are not your personal priorities is really unfair, and makes your motives suspect.
Roy Schestowitz
2008-09-18 16:52:59
bboissin
2008-09-18 17:03:34
> So, IMHO quite reasonable to assume that anyone doing kernel > development on Ubuntu will have to go through similar hoops to > get a backported patch accepted there - so the upstream gets its > patch, which might not have been written if someone wasn’t using > Ubuntu - but the patch isn’t contributed under Canonical’s name.
First Ubuntu kernel doesn't require a patch to be upstream for it to be shipped. If it did, people would say they are a very good citizen in the open source ecosystem, but that's not the case. Then if the patch wasn't contributed by Redhat or Canonical (like in you case where you submitted a patch), then it won't be counted in the stats from Greg. There's no reason why it should be. And just look at many major features (in Xorg or Gnome, etc), you will find they were sponsored by Redhat or Novell (network-manager, pulseaudio, compiz, etc).
The strength of Ubuntu is the intergration (and all the hacks they add to packages to make it work "better") but that doesn't help upstream at all it helps ubuntu.
(I confess I use Ubuntu, and I would really really love if they did more upstream work like redhat, I know that most if the new stuff I use comes from them).
Roy Schestowitz
2008-09-18 17:25:25
AlexH
2008-09-18 17:32:03
That said, I doubt if you broaden the cross-section you'd find Canonical's contributions leap hugely. If you look at https://wiki.ubuntu.com/Website/Content/UbuntuContributions there are relatively slim pickings, and some of which are already obsolete.
It's encouraging that they've been talking about doing some 'heavy lifting' recently, but I think it's fair to say that they haven't really done any yet.
Baby In The Bath Water
2008-09-18 19:24:35
If you take a look at the slides, Greg actually does mention his email address (gregkh@suse.de) on the opening slide.
Also, I'm not sure what you mean about Ubuntu's latest distro shipping a later version of the kernel than Novell's. Novell ships kernel 2.6.25 in the openSUSE 11 Gold Master.
See http://news.opensuse.org/2008/06/19/announcing-opensuse-110-gm/ for proof. Search for "kernel" which is in the "Under the hood" section.
Note: this comment was posted from Novell's headquarters.
Jose_X
2008-09-18 22:59:25
I'm not saying Greg doesn't support Novell's overall position or maybe really likes his environment. He is talented enough to get a job in other places, so this may very well be the case.
The point though is that we should be careful not to give a pass on things that may hurt wider FOSS simply because of some positives. We can shun one aspect and applaud the other.
Roy Schestowitz
2008-09-18 23:05:29
Canonical CTO bites back at Linux Plumbers Conference keynote
landofbind
2008-09-19 04:33:40
Today he is their defender. Why? Because the words of Mr. Kroah-Hartman were untrue? No! Because he works for Novell. And as we know the enemy of my enemy is my friend.
Why waste time taking into account the character and integrity of Mr. Kroah-Hartman when we can smear his name because he works for Novell. Novell that made a deal with Microsoft. Hell, he almost works for Microsoft. He must be a disciple of the devil.
So instead of pondering on what he said, we dismiss him because he works for Novell. There, the good job of the self-righteous.
Note: comment has been flagged for arriving from an incarnation of a known (eet), pseudonymous, forever-nymshifting, abusive Internet troll that posts from open proxies and relays around the world.
AlexH
2008-09-19 07:03:49
mdz's response is telling. "Oh, but we did send binutils a patch!". And then you see the patch. Hm.
I'm not sure why people don't see the danger in having large numbers of users with an organisation that doesn't actually do much free software development.
Roy Schestowitz
2008-09-19 07:24:41
AlexH
2008-09-19 07:56:29
Where am I saying "they are evil too?"
You keep accusing me of making this argument, and I've never stated that a single time.
Roy Schestowitz
2008-09-19 08:18:41
AlexH
2008-09-19 08:28:20
Roy Schestowitz
2008-09-19 08:30:12
AlexH
2008-09-19 08:43:21
You keep accusing me of making this argument, and I keep telling you that it's not an argument I'm making. Previously I thought you just didn't understand the points I was making; now I'm not sure you're even reading them.
And, I didn't insinuate anything about Ubuntu: the points I made were specifically about Canonical. I'll say it explicitly: they don't do much free software development. I've already pointed to their own page which lists the contributions, and compared to practically any other project they're not huge.
And I welcome their proposals to do some "heavy lifting". But it hasn't happened yet.
Roy Schestowitz
2008-09-19 08:54:13
No, but it's related to the pattern. Microsoft mishandles HTML? Point a finger at another vendor? Novell helps Microsoft with OOXML? Mention Sun. Mono is a risk? Claim that Java too is a risk...
They are not as large as Red Hat or Novell. Integration too enables many other derivative distros to reach a large audience (Mint, Parsix, Kiwi). Contribution comes in many forms, be it patches. promotion, legal work, documentation, support, etc.
If you're referring to Free software programming, I've done that too. Among other forms of contribution...
AlexH
2008-09-19 08:59:52
Sorry to ask the question again, but it's important: where, in this article, did I make that comparison?
I could go through the other points you raise and show you how they're false, but I addressed every single one of those false accusations in the previous articles so I see no need to repeat myself.
In terms of Canonical, I'm not comparing them to Red Hat or Novell, and neither is gkh, so your argument is an obvious straw man.
gkh's facts are not in dispute. If Canonical's contributions are large in some other area, then please show the data.
Roy Schestowitz
2008-09-19 09:14:37
See this comment (you disregard my answer). I described a general pattern, not made any statement about this post specifically.
You made a snide remark about the userbase, as it it's wrong to be a small company that's also successful.
Launchpad is your friend. BTW, Launchpad is likely to become AGPLv3-licensed.
AlexH
2008-09-19 09:27:17
So, "once again" where?
Roy Schestowitz
2008-09-19 09:32:20
AlexH
2008-09-19 09:39:44
Seriously. Your argument is "Yeah, but you're wrong, because I disagreed with you in this other completely unrelated article"? That's weak.
And you're right, the CSS story is a very nice example, because it shows clearly how far off base you are. Microsoft are implementing CSS 3 features, and like other vendors are prefixing them, like they are supposed to. You think this is "destroying the web", and when I showed you that Firefox do the same you think I'm defending them with an "equally evil" argument.
The fundamental flaw in your position is that you're starting from the assumption, in faith not fact, that Microsoft are wrong in this instance. I asked you five clear times to state what Microsoft should be doing instead, and you couldn't answer.
Roy Schestowitz
2008-09-19 09:44:18
AlexH
2008-09-19 09:49:08
a. retract your claim that I was making some kind of "they are evil too" argument if you're not able to support it with a citation
b. show us where Microsoft are wrong by prefixing CSS declarators.
And in general, please don't put forward arguments that you're unwilling or unable to support with any actual evidence.
Roy Schestowitz
2008-09-19 09:58:31
I was referring to a pattern and made no mistake. Java/Mono is one example of this.
This concern is not only mine:
http://www.noooxml.org/forum/t-89497/microsoft-hijacks-now-web-standards-and-the-w3c:m-fonts
Roy Schestowitz
2008-09-19 10:05:03
"Some missing numbers…
I dug up a few numbers that Greg missed.
* Worldwide Employees o Canonical:< 200 o Red Hat: ~2200 o Novell: ~4100 o IBM: 386,558"
AlexH
2008-09-19 10:12:26
Again, please retract this claim. If you think I'm making that argument elsewhere, put the comment elsewhere and I will refute it again. If you think I'm making it here, cite it. If you don't think I'm making it here, retract it.
Your NOOOXML link on the CSS story shows a basic misunderstanding. NOOOXML claim that Microsoft are extending CSS 2.1; this is pretty plainly false. The vendor extensions have been used on primarily CSS 3 properties.
Again, state what you think Microsoft should be doing. This is about the seventh time I've asked you this, and you still have literally no answer.
Dan O'Brian
2008-09-19 13:47:00
AlexH
2008-09-19 13:54:02
Perhaps he's just away working on his -roy- CSS extensions ?
BW
2008-09-19 14:14:10
Actually, nobody *requires* that patches be made in vanilla; that'd defeat the point of having distro-specific patches, since the distros do have a need to be at least a little unique. :) My experience is that large vendors will not accept third party patches without at least an LKML approval, if not an inclusion.
Also, I've had coworkers have patches "stolen" - we'll be working with RHEL or SLES in a bugzilla entry, post some code there, and then it'll appear on the LKML with a Red Hat or SUSE email address as the source. So I have to wonder if some number of those patches are not original work.
Also there's no legal requirement to contribute back to the origin point of a GPLed product, anyway; you have to ship your altered source, but only to the people who you've given the product. Now there's absolutely no reason some good Samaritan couldn't contribute diffs back to the vanilla, and in general patches do seem to end up in the upstream, but again, part of the value that RHEL and SLES add to the kernel is those custom patches.
@baby in the bath water
Is SLES 11 shipping yet? We're doing SLES 10 SP 2 here at the moment so I'm a little surprised to hear that SLES 11 is GMC already.
That said, enterprise Linux appears to work about the same as Windows and Solaris as far as customer adoption rate goes, so even with SLES 11 out and a 2.6.25 kernel available, I'm sure it'll be a year or two before it makes any real impact, at least on support work. Hell, we still have SLES 9 installations asking for support... in 32-bit mode.... on Xeons... >:/
BW
2008-09-19 15:55:45
I just recently (as of 11:02 AM) got the Novell PartnerNet announcement for beta 1 of SLES 11 - where did you get your information?
Roy Schestowitz
2008-09-19 16:18:49
SLED 11 is not even being /discussed/ in public by Novell. It's far from a finished product and I think it'll be based on the x.1 release, as usual. That's what one of the OpenSUSE guys told me.
Anonymous
2008-09-19 16:53:12
Baby In The Bath Water
2008-09-19 18:57:22
Note: this comment was posted from Novell's headquarters.
Jose_X
2008-09-19 20:00:30
No. I expressed no opinion at all on the topic of this article. I truly just made a side comment.
>> Your NOOOXML link on the CSS story shows a basic misunderstanding. NOOOXML claim that Microsoft are extending CSS 2.1; this is pretty plainly false. The vendor extensions have been used on primarily CSS 3 properties. >> Again, state what you think Microsoft should be doing.
I read over the blog (quickly), and I think it clearly said that they were extending CSS 2.1; however, the blogger also adds that this is legal.
It appears to me that Microsoft is doing what they should be doing (I'm not going to bother to dig out the spec or to reread the blog if no one disagrees), at least to the extent that they would be adding extensions and wrt the namespace prefixing on those extensions.
Here is the deal though. Extensions allow for legalized E3 from a monopolist. There is no getting around that. Extensions can be useful to any vendor/product/user, and extensions can also be used to lock users in. ODF has this same issue, as likely do many other standards.
The problem isn't the extension framework. The problem is Microsoft's business plan.
They can E3 legitimately as concerns the standards. Users need to shun closed source monopoly product regardless of the standard sticker on the box. These standards won't save you. Misconception about this will lead people to give Microsoft a pass when they fulfill their standards obligations.
AlexH
2008-09-19 20:17:43
You entirely misunderstand why they're using a vendor prefix. Look at the Microsoft page, and the CSS3 declarations, and compare with what Firefox are doing. This isn't about CSS 2.1.
Roy Schestowitz
2008-09-19 20:24:59
AlexH
2008-09-19 20:28:16
Now that you're around, do you mind stating what Microsoft should be doing now?
BW
2008-09-19 21:29:55
Ah, I missed that little change-up you threw there - which means, you've probably missed my entire point, and probably don't work in the enterprise software business.
There is a HUGE difference between SLES and OpenSUSE. The former is an enterprise-level release, which customers use in situations where they need a well-defined and stable platform, as their business depends on their systems being up 100% of the time and not subject to frequent updates. It also means that the vendor will have support available for several years (hence my complaint about SLES 9). The latter is a community release which comes with no guarantee of support and no promise that the day's Yast update won't wreck your system beyond repair.
The vital difference here is that Ubuntu LTS releases are treated as enterprise releases with a 3-year support lifespan - and so, with Canonical's small number of employees, this means that they spend the majority of their time keeping their existing release working and not doing front-end development or vanilla bug-fixing.
Compare that to Red Hat or Novell, with over 10 and 20 times the employee base respectively, and very high-profile connections to hardware developers (like IBM, Intel, AMD, LSI, to name a few), and there is no possible way that little Canonical could contribute nearly the same number of patches as either.
Does that make Canonical's contributions worth less than Red Hat or Novell? No.
Does that mean that anyone with less contributions than Red Hat or Novell is, in effect, not giving back to the community and should therefore be treated dismissively? No.
Should Gary go and put together a similar presentation for say, Slackware, who didn't even make the top 10 list? Hell no!
I view what Gary said as insulting because it essentially draws a line in the sand and says "if you're this popular, you must contribute at least X number of patches to be worthy of praise" - when ALL contributions should be considered worthwhile, something which is at the core of open source development. I think Roy probably would agree with me.
Baby In The Bath Water
2008-09-19 22:41:07
You mentioned Ubuntu ships the latest kernel - I didn't realize you were comparing "Enterprise Server" editions (which, afaik, Ubuntu does not have).
It's also a bit unfair seeing as how the release schedules of Ubuntu, Red Hat, and Novell are not in sync. They each (afaik) release a new "Enterprise Edition" every 3 years - if you want to compare apples to apples, then the best you could do is compare roughly-equivalent releases.
E.g. Ubuntu 8.04 vs SLE[D,S] 11 vs RHEL [whatever]
Obviously it's unfair to compare Ubuntu 8.04 released just 5 months ago to something released 3 years ago.
Likewise it's probably unfair (albeit much less so) to compare SLE[D,S] 11 (when it is released) to Ubuntu 8.04 since Ubuntu will be older (~9 months older? I have no idea what the release schedule for SLES 11 is).
I'm sure you get my point.
As far as the criticisms of Ubuntu go, they are no worse than this site does to Novell. Roy and his supporters often attack Novell claiming they make little-to-no contributions to FOSS (which is quite clearly untrue).
Personally, I see both sides of the argument.
Canonical are seen as a bit of a leech by the developer community (I've seen these complaints from sources other than Greg KH) because they take a lot and give back little. The difference between them and Slackware is that Slackware is not commercial and is thus deemed more acceptable by the developer community. Debian, another non-commercial distro, gets extra brownie points for contributing a lot back to the kernel, etc.
On the other hand, as Roy pointed out, they do make a very successful distro that "the average joe" can use which arguably attracts more people to Linux/FOSS, thus helping Linux/FOSS gain more acceptance.
I'm the type of developer that licenses his work as MIT/X11 whenever I can with the desire to allow anyone to use it, proprietary vendors included, and don't particularly care about being compensated nor inflicting my personal beliefs on anyone else.
I mention this because in my eyes, Canonical isn't doing anything wrong in this particular case (but I understand the criticism).
Note: this comment was posted from Novell's headquarters.
Jose_X
2008-09-20 00:47:07
If this is your more precise reply http://boycottnovell.com/2008/09/13/microsoft-admitted-mono-trap/#comment-24351 , I replied in that thread a few posts below that one.
Roy Schestowitz
2008-09-20 01:28:46
http://www.phoronix.com/scan.php?page=article&item=ubuntu_bulletproofx_2&num=1
"Introduced in Ubuntu 7.10 was a feature known as BulletProofX, which provides a fail-safe mode that is by default used when the X server fails to properly initialize. In this original implementation, it would default back to using the VESA display driver with 256 colors and then proceed to run the displayconfig-gtk utility. While this is nice for the end-user as it keeps them from touching a terminal to debug an X server problem, for experienced users it inhibits them from easily debugging the problem."
Victor Soliz
2008-09-20 02:36:14
Regardless, number of commits won't change the fact of Novell continuous attempts at boycotting other Linux distributors using FUD, the horrid attempts to make everyone grow a Mono dependency, the MS technology advocation, including even the infamous OOXML, etc, etc, etc. They add a lot of patches, however, unlike what apologists would think, that doesn't give them any right to do what they do.
AlexH
2008-09-20 08:44:00
I'm not sure the criticism was purely the number of commits; I think it was that the number was very close to zero.
Roy Schestowitz
2008-09-20 08:56:03
Either way, Slated posted a very interesting message some minutes ago (USENET MSG-ID <ctcd5eq5-093.ln1@sky.matrix>). It shows another perspective, so I add it here with his permission.
AFAICT Canonical's biggest contribution to GNU/Linux has been to indoctrinate millions of noobs into believing that Free Software advocates are "fundamentalists" (Shuttleworth's own word, ref: the podcast interview about the end of Gobuntu); that "trivial" concepts like Freedom are less important than convenience (ref: Gobuntu mailing list posts about Mozilla€® trademarks); and that it's OK to stab Free Software developers in the back by marginalising their work in favour of pandering to the demands of Intellectual Monopolists (ref: Ubuntu Remix codecs).
The result is an army of clueless Ubuntu fanboys who genuinely consider Ubuntu as being somehow distinct from GNU/Linux ("it's the 'Ubuntu' operating system"); have nothing but contempt for the FSF, the GPL, Stallman and the principles of Freedom; think like Windows users; talk like Windows users; act like Windows users; and aspire to the same twisted goals and follow the same worthless paradigms as Windows users. In short, they are Windows users ... who got lost, and tried to turn GNU/Linux into Microsoft Slopware, all with Shuttleworth's hot and eager help, of course. Not that many of them have the skills or enthusiasm to "turn" anything into anything else, since they are first; last and always "users" (roughly translated as "leechers, whiners and fanboys"), which may account for why comparatively little is contributed to Free Software from the Ubuntu sheep-pen.
Now it seems that Shuttleworth has a hardon for Macs, so I suppose the next generation of Ubuntu converts will be Mac switchers, all demanding "killer apps" with "i" appended to their names (why does this make me think of "There's no 'I' in 'Team'"?). I fully expect some distant version of Ubuntu to be called "iBuntu"; they'll drop support for mice with more than one button; and they'll obfuscate so much of the OS for the sake of "simplicity" that your only interface for pretty much everything will be a single button that reads "OK" ... and even that single "choice" will eventually be depreciated in favour of a completely non-interactive screen that just runs "Get a Buntu" adverts ("Hi, I'm a Buntu. Hi, I'm a Buntu too").
Canonical has not helped spread the adoption of GNU/Linux, they've helped spread the adoption of the *bastard son of Linux*; a cancer -ridden mutation spawned by inbreeding between Windows developers; Microsoft fanboys and a bunch of naive children.
Shuttleworth is no more a "Linux guy" than Novell is a "Linux company", they're both just opportunists riding the wave. Novell may be 25 years old, but they've only spent the last 5 of those years as a Linux company (after acquiring Ximian from Microsoft fanboys de Icaza and Friedman, who's biggest contributions to GNU/Linux are patent-encumbered clones of Microsoft technology). And the last 2 of those 5 years have been spent as Microsoft bum-boys, in a protection racket designed to give Novell "exclusive privileges" at the expense of the Free Software community.
Also note that SUSE != Novell. I have nothing against the SUSE distro, I just happen to have nothing but contempt for Free Software sellouts. Buried beneath all that management-driven Microsoft fanboyism, is a truly excellent German Linux distro desperate to break Free. Again. Unlike Novell, SuSE was distributing Linux 14 years ago (a few months before Red Hat).
When I think "Novell", I don't think "Linux", I think "Netware", and being shafted by Microsoft ... twice. I also think about their efforts to undermine Free Software with Poisonware like Mono and Moonlight, and about their attempts to co-opt OpenOffice.org with yet more Novell and Microsoft "exclusivity", to conveniently shut-down the threat to their pal Microsoft's ailing cash cow.
The considerable contributions SUSE developers make to Free Software is completely undermined by this intolerable situation, and also brings into question the integrity of those contributions, rightly or wrongly. I'm not claiming that every SUSE user and developer is a bad guy (much like my observations of Ubuntu are just generalisations), but they are currently standing in the heart of enemy territory, and that simply doesn't look good for any of them.
Mainly I can't help but think that people who would climb into bed with the self-declared enemy of Free Software are the worst traitors in our community. How any of them can even dare to show face anywhere beats me, as their new "partner" denounces Linux as a "cancer"; perverts the ISO Standards process with bribery and smear campaigns; sabotages charities like the OLPC; corrupts Nigerian education suppliers with bribes to oust Mandriva from signed contracts; and holds the whole world to ransom with a monopoly supported by a racketeering operation with OEMs, that seeks to exclude all competition using clandestine MoUs that are protected from public scrutiny by equally corrupt laws.
As for Ubuntu, AFAIAC that's only ever been one thing ... "Proprietary Debian", a.k.a. "Windows Remix". But then a "fundamentalist" like me is bound to say something like that, eh?
Being a Free Software advocate no more makes me a "fundamentalist" than dissent against Microsoft's racketeering operation makes me a "hater". It's not wrong to denounce crime and bad ethics, and it's not wrong to support Freedom. What is wrong is to marginalise those who do take that position as "fundamentalist"; "loons"; "zealots" and "haters", especially when those who are doing that marginalising are traitors to the cause they pretend to support (a.k.a. "pragmatists"). Such people are actually more dangerous than outright Microsoft Evangelists, because they are close to the source, and have a background that ostensibly makes them seem "reasonable" and "balanced", but in reality they are poisoning the Well that the rest of us must drink from to survive. Not that these "pragmatists" need to worry about poisoned Wells, since they have endless supplies of Microsoft-branded Kool-Aid€® to keep them going.
Supporting Free Software is about more than just supporting the principles of Freedom, it is a common courtesy and a mark of respect to the countless numbers of developers who gave their time and effort Freely and willingly ... for our benefit. What Shuttleworth fails to understand is that every time he opens his gob and denounces Free Software advocacy as "fundamentalism", he is spitting in the face of every Free Software contributor and the *vast bulk* of the codebase that forms "his" distribution ... that he then promotes as the One True Linux€® and proceeds to poison with proprietary Slopware, whilst brainwashing his flock into mocking Stallman and believing that "becoming OSX/Windows/Anything-but-Linux" should be their wet-dream.
Thanks, but I'll pass.
Dan O'Brian
2008-09-20 12:51:59
But you attack Greg KH for criticizing Ubuntu while your friend attacks them much more harshly? And you support [H]omer's attacks, but dismiss Greg's? At least Greg's criticism comes with real numbers to back his claim. [H]omer's attack is just ranting.
Dan O'Brian
2008-09-20 12:55:42
Roy Schestowitz
2008-09-20 12:58:39
No, I don't agree with most of this and I refuted/rebutted in USENET.
Dan O'Brian
2008-09-20 13:14:38
(I don't read USENET nor would I know where to find the message/thread you are referring to).
Roy Schestowitz
2008-09-20 15:18:35
Another important player which I believe gives very little to the kernel is Mandriva. It's probably the best distribution out there at the moment (2008.1 is what I use on my main box). This brings desktop users to the kernel. That can't harm it, can it?
Lastly, speaking of so-called 'market share' (or installbase rather), there is a lot of deception out there.The Gartners or the world are part of this... the same Gartner that does a lot of business with Microsoft, accepts personal investments from Bill Gates and says that only 4% if the desktops/laptops out there run GNU/Linux while GNU/Linux is used on servers a lot more than people realise (never mind devices and supercomputers).
Ubuntu gives us many new users. It's a blessing, not a curse. It's an expansion of Free software userbase, not distro cannibalisation (although that too is debatable).
bboissin
2008-09-20 15:29:25
AlexH
2008-09-21 08:10:24
Just out of interest, what is this "a lot" it does for the free desktop?
Roy Schestowitz
2008-09-21 08:23:47
AlexH
2008-09-21 08:55:50
I'm sure they do a fair amount for the popularity of the OS, but a decent advertising campaign would do that.
Roy Schestowitz
2008-09-21 09:09:05
Ubuntu giving back to Debian: facts and numbers!
The Heron takes flight
Debian and Ubuntu
AlexH
2008-09-21 09:23:26
As an example, the Utnubu project seems to be mostly dead: how many Ubuntu patches are getting into Debian?
It's not like this criticism is new; in fact it has been raised many times by many different people.
Roy Schestowitz
2008-09-21 09:27:52
What distribution do you use anyway?
AlexH
2008-09-21 09:39:19
I haven't seen a rebuttal to that based on facts. Some of their patches go upstream, but not many of them. This is well documented.
Roy Schestowitz
2008-09-21 09:41:47
AlexH
2008-09-21 09:59:45
What is discouraging about their recent statement about "heavy lifting" is that Mark S expressly said they were going to fork the projects, develop on Launchpad, and then try to merge stuff back upstream.
I left some comments on his blog asking to reconsider that, but I don't hold out much hope. I would compare that method to the stuff that Novell did with the GNOME menu: it's a nice design, the search and stuff is a nice touch, but when you don't develop upstream it's very difficult to push it as a "finished product". I hope Novell learned from that experience, and I hope Canonical also learn from it.
AlexH
2008-09-21 11:00:46
Perhaps Greg could have made the same points in a less confrontational way, but it looks like things will get even better anyway.
Roy Schestowitz
2008-09-21 11:35:12
No, I can't find it; not quickly anyway. While searching, however, I stumbled upon:
Lucas Nussbaum: Ubuntu information on the Debian Package Tracking System and the Developer Packages Overview
Canonical Joins The Linux Foundation
Anyway, the one I was looking for pointed out what Greg did. It's about 2 months old and the author did a search on patches by E-mail suffix. I believe he then defended Ubuntu/Canonical. I submitted this to Digg at the time.
AlexH
2008-09-21 13:32:22
It's difficult to get decent numbers on this, but he's in the right ballpark.
Roy Schestowitz
2008-09-21 15:52:33
AlexH
2008-09-21 16:19:53
Roy Schestowitz
2008-09-21 16:21:56
http://www.linux-foundation.org/weblogs/amanda/2008/09/19/free-riders-canonical-and-greg-kh/ http://mdzlog.wordpress.com/2008/09/17/greg-kh-linux-ecosystem/ http://dustinkirkland.wordpress.com/2008/09/18/whats-behind-gregkhs-latest-rant/
AlexH
2008-09-21 16:23:39
Roy Schestowitz
2008-09-21 16:46:14
Dan O'Brian
2008-09-21 17:25:53
Roy Schestowitz
2008-09-21 17:42:09
AlexH
2008-09-21 17:50:38
Roy Schestowitz
2008-09-21 17:58:50
"As Matt Zimmerman discussed, Greg’s “Linux ecosystem” seems a bit unfairly limited to the kernel, gcc, and binutils, and neglects a wider macrocosm of Ubuntu’s contributions to the Linux, free, and open source space. Canonical and Ubuntu actively contribute to GNOME and KDE, as well as dozens of other open source projects (e.g., I’m co-maintainer of the upstream eCryptfs project and have contributed considerable code there on Canonical’s dime)."
AlexH
2008-09-21 18:02:34
Greg's presentation explicitly addressed that point when he made it. It was the "Linux Plumber's Conference", so he looked at the "Linux plumbing".
That doesn't mean that Canonical contribute widely outside those applications he looked at. Are the numbers better? Has anyone actually looked?
For example, I found it quite depressing to see that the "Netbook interface" isn't under the aegis of GNOME.
Roy Schestowitz
2008-09-21 18:05:52
This was not disputed. But he chose them as a target based on a narrower scope.
It's akin to Microsoft daemonising Linux by telling stories about Hans Reiser; him and him alone (and yes, its mouthpieces do exactly that).
AlexH
2008-09-21 18:19:24
You think he should be talking about issues outside the "Linux plumbing"?
I think in general it's usual - and indeed preferred - that keynote speakers "narrow" their "scope" to the topic of the conference.
Roy Schestowitz
2008-09-21 18:24:23
AlexH
2008-09-21 18:27:33
You still keep talking about their contributions elsewhere; which contributions are these? If they're sending lots of patches to GNOME/KDE I'd be hugely impressed, but I really suspect that they're not. It would be nice to see some actual numbers before speculating.
Roy Schestowitz
2008-09-21 18:31:19
Roy Schestowitz
2008-09-21 18:31:46
AlexH
2008-09-21 18:35:35
I suspect Canonical don't contribute much to apps outside those Greg looked at, except for a couple of exceptions like bzr. But, I'm not going to assert that, I don't know and neither does Greg.
The argument also isn't about the size of the contribution, so your request for normalisation is also irrelevant, as you ought to know as a researcher.
Roy Schestowitz
2008-09-21 18:38:03
AlexH
2008-09-21 18:43:34
Dan O'Brian
2008-09-21 18:45:41
Roy Schestowitz
2008-09-21 18:54:01
http://www.forwardcamegrendel.org/decline-gentoo-linux
RyanT
2008-09-21 18:59:27
We're not comparing distros, we're comparing companies.
In this case, yes, it is unfair of a much larger corporation like Novell to pick on a comparatively small child like Canonical. There's already been rebuttals in several places, per the links Roy posted and others.
Not to mention clarification on Canonical (and Ubuntu's) position. They seem themselves more as integration folks, not to mention the recent announcement of more manpower to go towards mostly GUI related areas.
All this boils down to is an Novell employee wanting to whip it out and see how big it is compared to others.
http://news.cnet.com/8301-13505_3-10046449-16.html
Dan O'Brian
2008-09-21 19:33:20
The issues that Greg talked about still stand.
Roy Schestowitz
2008-09-21 19:42:55
AlexH
2008-09-21 19:54:06
Roy Schestowitz
2008-09-21 21:27:30
Dan O'Brian
2008-09-21 21:12:13
330 people is not much more than Ubuntu.
Dan O'Brian
2008-09-21 22:51:11
Dan O'Brian
2008-09-21 22:46:40
Roy Schestowitz
2008-09-21 22:48:20
Roy Schestowitz
2008-09-21 22:53:53
Dan O'Brian
2008-09-21 22:57:43
Dan O'Brian
2008-09-21 22:58:07
Roy Schestowitz
2008-09-21 23:03:11
AlexH
2008-09-22 06:51:17
What they are doing is working on things like new widget layout so that dialog and UI doesn't have to be created in code. And making it faster to start up.
Have you actually looked at the code that Novell sends upstream into OOo? Why do you assume it's just Mono and OOXML?
Roy Schestowitz
2008-09-22 07:37:19
AlexH
2008-09-22 07:55:34
Roy Schestowitz
2008-09-22 08:00:36
AlexH
2008-09-22 08:11:43
Roy Schestowitz
2008-09-22 08:21:19
AlexH
2008-09-22 08:49:51
You're suggesting that had Novell not "supported" OOXML, that Microsoft would have never put it into Office 2007 and we'd never have to deal with it?
I find that unlikely in the extreme, given that Office 2003 was already supporting XML formats....
storagemonster
2008-09-22 09:06:11
Roy Schestowitz
2008-09-22 09:13:59
AlexH
2008-09-22 09:16:58
I don't see any logical connection whatsoever.
Roy Schestowitz
2008-09-22 09:19:12
AlexH
2008-09-22 09:20:22
Which argument are you trying to make? You can't argue two conflicting points of view simultaneously.
Roy Schestowitz
2008-09-22 09:21:55
AlexH
2008-09-22 09:26:09
Are you suggesting that if Novell hadn't participated in TC45, that Sun wouldn't be implementing it now?
Roy Schestowitz
2008-09-22 09:37:16
AlexH
2008-09-22 09:46:59
Roy Schestowitz
2008-09-22 09:48:53
AlexH
2008-09-22 09:50:56
stevetheFLY
2008-09-22 10:24:00
Note: comment has been flagged for arriving from a possible incarnation of a known (eet), pseudonymous, forever-nymshifting, abusive Internet troll that posts from open proxies and relays around the world.
Ian
2008-09-23 01:36:29
"In this case, yes, it is unfair of a much larger corporation like Novell to pick on a comparatively small child like Canonical. There’s already been rebuttals in several places, per the links Roy posted and others"
GregKH is a Novell employee, not the entire company. He's a single developer. He doesn't drive, at least directly, the vision and goals of a company. The only mentions of Novell are in the slides, and they don't even paint Novell as the golden child of committed patches. They actually give a nod to Novell's real competitor and market leader in the enterprise linux realm(red hat). Moreover, it's about Linux and some of the related tool chain, not open source or the linux desktop in general.
Sometimes there's just no conspiracy there.
Roy Schestowitz
2008-09-23 08:29:29
Oops! Ubuntu IS gearing up for more kernel contribution
"Ubuntu is on the verge of fully participating in the Kernel Oops project run by Intel's Arjan van de Ven (i've interviewed him before, great guy and an awesome project). Kernel Oops is an incredibly valuable effort that tracks 'oopses' on Linux and provides information so kernel developers can fix bugs. So far Kernel oops has been part of the default installation on Red Hat's Fedora and is available to Novell OpenSUSE users as well..it soon may beavailable by default to Ubuntu users too."
http://blog.internetnews.com/skerner/2008/09/oops-ubuntu-is-gearing-up-for.html
Dan O'Brian
2008-09-23 13:11:56
Grats. </sarcasm>
That doesn't change anything, they should have been doing that for a while now.
Roy Schestowitz
2008-09-23 13:20:34
Dan O'Brian
2008-09-23 14:19:56
Not to mention it wouldn't have been difficult for them to ship the kernel-oops reporter even if it was only a handful of developers. It's not like Canonical has to write their own tool.
Roy Schestowitz
2008-09-23 16:41:05
[Again, I'm not sure.]
Roy Schestowitz
2008-09-23 19:36:06
http://ostatic.com/173422-blog/the-open-source-contributions-of-six-blind-men-and-an-elephant
"Herein lies the problem.
"Kroah-Hartman works on the kernel, the "Linux ecosystem." He was delivering an address at a conference geared to a particular audience -- people who work on the core of the Linux system. In this light, his arguments appear completely valid. He couldn't rightly comment on other areas of development, as they weren't in the scope of his experience, or the conference's focus. This is also well within reason.
RyanT
2008-09-23 21:50:09
I never said there was. I was pointing out that in measuring contributions, you must remember company sizes and priorities.
It's been pointed out in many places - the definition of ecosystem being hazy, ignoring Mark Shuttleworths own clarifications about where Ubuntu and Canonical are specifically aiming, and writing off recent announcements of upped development support in several areas.
I always remember one of the main arguments of open source being "everyone can work on what they're interested in!" before making the switch, amongst many others to the same effect. Then you get constant crap like this where people start cat fights because not everyone is working on what they're interested in.
Roy Schestowitz
2008-09-23 21:59:44
Maybe Canonical isn't feeling 'itchy' in the kernel dept.