| |||||||||||||
|
| |||||||||||||
Vignatti: X Census (for 1.9)Vignatti: X Census (for 1.9)Posted Sep 2, 2010 15:41 UTC (Thu) by patrick_g (subscriber, #44470)Parent article: Vignatti: X Census (for 1.9) Only one line changed for Canonical ? Oh my God, it's even worse than for the kernel of for Gnome.
(Log in to post comments)
Vignatti: X Census (for 1.9) Posted Sep 2, 2010 15:58 UTC (Thu) by gowen (guest, #23914) [Link]
Jef Spaleta called - he wants his schtick back.
Vignatti: X Census (for 1.9) Posted Sep 2, 2010 16:37 UTC (Thu) by ewan (subscriber, #5533) [Link]
It's not just Jef. Canonical's approach to this issue has been utterly shambolic and they thoroughly deserve most of the flak they get.
Vignatti: X Census (for 1.9) Posted Sep 2, 2010 17:02 UTC (Thu) by patrick_g (subscriber, #44470) [Link]
To be fair I expect better statistics for Canonical concerning the 2.6.36 kernel (because of AppArmor integration).
Vignatti: X Census (for 1.9) Posted Sep 3, 2010 8:22 UTC (Fri) by AlexHudson (guest, #41828) [Link]
What proportion of that do you attribute that to Canonical though (rhetorically)?
I don't mean to dismiss the work they're doing, because if they can complete the integration that would be excellent, but from an uninformed observer's point of view probably 90% of the work had been done for them.
Canonical are supposed to be doing all this work with Dell and other OEMs, I would have expected some steady trickle of hardware-based patches - not necessarily whole drivers - and that doesn't seem to be the case. That seems really, really odd.
Vignatti: X Census (for 1.9) Posted Sep 3, 2010 10:22 UTC (Fri) by hppnq (subscriber, #14462) [Link] Doesn't it make slightly more sense to expect hardware-based kernel contributions from Dell?Speaking as someone who has been following Free Software development long enough to remember the time when it was not yet the product of Corporate America -- but of small outfits with obscure names such as Red Hat or Slackware, until IBM announced it would invest a cool billion -- the discussion about who should contribute what has always seemed a bit strange. It's not as if the many companies involved in Free Software development invest resources that they would otherwise have spent on their core business: these contributions are their core business.
Vignatti: X Census (for 1.9) Posted Sep 3, 2010 10:31 UTC (Fri) by drag (subscriber, #31333) [Link]
> It's not as if the many companies involved in Free Software development invest resources that they would otherwise have spent on their core business: these contributions are their core business.
Yes. This is the important thing to remember about professional contributions:
They are, on the majority of cases, not done through kindness or correctness or being "socially responsible". They are done through selfish purposes.. these companies stand to benefit financially to contributions done for the kernel. Whether they are fixes done to correct bugs, improve the attractiveness of their hardware or support or anything.. they stand to benefit from contributions and efforts made to open source/Linux.
This is what makes it so wonderful. :)
Vignatti: X Census (for 1.9) Posted Sep 3, 2010 13:21 UTC (Fri) by hppnq (subscriber, #14462) [Link] This is what makes it so wonderful. :) Exactly! But difficult to behold in between all those stakes. ;-)
Vignatti: X Census (for 1.9) Posted Sep 3, 2010 11:26 UTC (Fri) by AlexHudson (guest, #41828) [Link]
It would make sense if Dell were the ones doing the integration, for sure, but then I don't know why they would have hired Canonical if they already had that level of expertise and commitment. Presumably the win for people using them as an OEM platform is that they handle the integration and make sure stuff works?
Vignatti: X Census (for 1.9) Posted Sep 3, 2010 13:14 UTC (Fri) by hppnq (subscriber, #14462) [Link] I think we agree that Ubuntu -- or any distribution -- making sure their stuff works on Dell systems means that it's a good idea for them to make sure the relevant Dell stuff works. But after that, you lost me.I would not expect any other party than Dell to support their own hardware. Nor would I expect Dell to support the Ubuntu distribution. I would not even expect Dell to support the Ubuntu effort to get Dell stuff working in their particular distribution. But I would expect some form of cooperation between the many parties involved in the whole process, in which their efforts effectively add up to something useful. Something like this, for instance, or this.
Vignatti: X Census (for 1.9) Posted Sep 3, 2010 10:04 UTC (Fri) by error27 (subscriber, #8346) [Link]
The lines changed count doesn't always reflect the value of the contribution. AppArmor isn't a lot of lines.
git log -p v2.6.35.. --author=Johansen | diffstat | tail -n 1
So John Johansen will show up as an individual contributor but probably Canonical won't show on the company contribution list.
But really I'd emphasize again that lines of code isn't a an accurate way to measure contributions.
Vignatti: X Census (for 1.9) Posted Sep 2, 2010 17:06 UTC (Thu) by boudewijn (subscriber, #14185) [Link]
And you know what? With all the whining about how Canonical doesn't contribute to the kernel, to Gnome, to X11 -- it's still the most polished, usable and pretty linux distribution of all.
And I say this as a hard-core KDE developer and OpenSUSE user. Maybe the "upstream" projects should take note and start to pro-actively re-integrate some of their work?
Vignatti: X Census (for 1.9) Posted Sep 2, 2010 17:09 UTC (Thu) by ewan (subscriber, #5533) [Link]
If the other distro companies stopped what they're doing and concentrated on 'polish' the way Canonical do there'd be no new stuff to polish. Someone's got to actually do some development.
Vignatti: X Census (for 1.9) Posted Sep 2, 2010 17:23 UTC (Thu) by boudewijn (subscriber, #14185) [Link]
And if they don't they will never have something that's really polished and nice to use... Or maybe it's time to recognize that providing polish is development too -- you know, the last 10% that takes 90% of the effort and all that sort of rot.
And speaking for myself, again as a KDE developer and OpenSUSE user -- I am the guy who is producing new stuff anyway. Not Canonical, not RedHat, not Novell. Of course, I'm only working on an application... Not something hugely important.
Vignatti: X Census (for 1.9) Posted Sep 2, 2010 17:28 UTC (Thu) by hadess (subscriber, #24252) [Link]
What's your app? So I can see whether you're the one producing, and certainly not me (apparently).
Vignatti: X Census (for 1.9) Posted Sep 2, 2010 17:55 UTC (Thu) by Sho (subscriber, #8956) [Link]
Boudewijn is the lead dude for Krita (http://krita.org/).
Vignatti: X Census (for 1.9) Posted Sep 2, 2010 20:08 UTC (Thu) by tcourbon (subscriber, #60669) [Link]
And that's not something I would call "not hugely important". He's way too modest.
Vignatti: X Census (for 1.9) Posted Sep 2, 2010 20:06 UTC (Thu) by boudewijn (subscriber, #14185) [Link]
As Sho has said, I'm the maintainer of Krita. Though I'm certainly not the most productive developer on Krita -- there's a great team of wonderful people doing amazing stuff I wouldn't be able to do on my own -- I do feel I have over the past seven years gained something of an appreciation on how difficult it is to make something that is actually polished and user-ready.
For one thing, it's much harder than cranking out new stuff.
Vignatti: X Census (for 1.9) Posted Sep 2, 2010 23:02 UTC (Thu) by boog (subscriber, #30882) [Link]
If Ubuntu wasn't useful, people wouldn't be using it. But it's by far the most popular distribution!
As has been noted elsewhere, Ubuntu are much smaller than, for instance, RedHat, and they are bringing in many users who require a lot of help. They seem to be quite busy.
Also, as far as I am aware, the whole enterprise is still costing Shuttleworth serious money. Maybe he can spend his money making a difference as he wishes? I think that creating a beginner-friendly Linux distribution is a very worthy goal. It's good if the other distributions feel some pressure on this front.
Vignatti: X Census (for 1.9) Posted Sep 2, 2010 23:46 UTC (Thu) by ewan (subscriber, #5533) [Link]
Nobody said it wasn't useful, and indeed, if it wasn't then I rather think that people wouldn't care so much. The problem is that the other distro companies (most notably, but far from exclusively, Red Hat) put a lot of work into the upstream projects (the kernel, Gnome, X.....) and that benefits their users, and everyone else's users too, including Canonical's. This is all good.
Canonical put their effort into making a nice, polished, beginner-friendly system, which is also all good, but they do the work in their own forks, not directly upstream, which makes it much harder than it needs to be for that work to benefit everyone else. Essentially, they're all take and no give, and that is what is not good.
Canonical then see fit to add insult to injury by coming up with nonsense like claiming that they do work upstream, but their desktop upstream is Ayatana (coincidentally an in-house Canonical project) rather than Gnome. It's completely transparent, and there's really no good reason for it.
Vignatti: X Census (for 1.9) Posted Sep 3, 2010 16:26 UTC (Fri) by chromatic (guest, #26207) [Link] Essentially, they're all take and no give.... Are you claiming that Canonical isn't increasing the amount and reach of software freedom in the world? Are you doing so based on SLOC counts? That argument seems shaky.
Vignatti: X Census (for 1.9) Posted Sep 3, 2010 17:43 UTC (Fri) by ewan (subscriber, #5533) [Link] Are you claiming that Canonical isn't increasing the amount and reach of software freedom in the world?No, I'm claiming that they're contributing less code to the common pool of projects that all(/most) distributions draw from, and that their work therefore does not benefit (for example) Red Hat's users in the same way that Red Hat's work benefits Canonical's users. Further, I'd suggest that there's no need for them to behave that way - doing development collaboratively and upstream demonstrably works, and Canonical could do theirs that way too, and as well as helping everyone else it would lead to a more sustainable end result for people using Ubuntu directly, since history suggests that maintaining divergent forks is not easy to do well. This is not an anti-Canonical position; I'd like to see more use, in more places of more Canonical work, not have it relegated to a single distribution.
Vignatti: X Census (for 1.9) Posted Sep 3, 2010 5:36 UTC (Fri) by seyman (subscriber, #1172) [Link]
> it's still the most polished, usable and pretty linux distribution of all.
Does anyone have any data to back this up (apart from kernel crashes and pulseaudio not working as well as on other distributions)?
Vignatti: X Census (for 1.9) Posted Sep 3, 2010 8:13 UTC (Fri) by seyman (subscriber, #1172) [Link]
> Popularity is your best measure for that.
Are you seriously suggesting that Google Trends is the best dataset to support the claim "[Ubuntu is] still the most polished, usable and pretty linux distribution of all"?
Vignatti: X Census (for 1.9) Posted Sep 4, 2010 16:32 UTC (Sat) by madscientist (subscriber, #16861) [Link] Since "usable", and "pretty" are completely subjective, yes, I'd say that popularity is a perfectly legitimate metric to use to support it. How else would you measure those?You can probably make an argument about "polished", by examining bugs reported etc., but that's a hard metric since you have to normalize it. Of course you may feel that Google Trends is not the best way to measure popularity, but again.... how else would you measure it?
Vignatti: X Census (for 1.9) Posted Sep 5, 2010 10:58 UTC (Sun) by seyman (subscriber, #1172) [Link]
> Since "usable", and "pretty" are completely subjective,
I wouldn't call "usable" completely subjective. It's far less so than "pretty", IMHO.
> yes, I'd say that popularity is a perfectly legitimate metric to use to support it.
Impossible, popularity is a non-linear metric. Besides, popularity doesn't solely measure usability and prettiness, it also measures perception of stability, perception of security, marketing, ...
> How else would you measure those?
There are usability experts who can measure it. Prettiness is a lost cause, IMHO.
> Of course you may feel that Google Trends is not the best way to measure popularity, but again.... how else would you measure it?
Anything that doesn't take into account actual use is a non-starter, that's for sure.
Vignatti: X Census (for 1.9) Posted Sep 6, 2010 23:14 UTC (Mon) by nevyn (subscriber, #33129) [Link]
> Since "usable", and "pretty" are completely subjective, yes, I'd say that popularity is a
> perfectly legitimate metric to use to support it.
By that rationale Windows 7 is more usable and pretty than Mac OS X.
My personal opinion is that Ubuntu is very popular due to the free CD program they ran and the network affects that followed from it. Technically they also managed to fork Debian well, certainly much better than anyone had done previously (and more than a few tried), which allowed them to keep users by making moving to the next Ubuntu easier than trying anything else (thus. increasing the network affect).
It's kind of sad that even LWN give them such a free pass, but maybe that will change in another 5 years. IMO they will either destroy themselves, destroy Linux, or possibly make become a great leading Linux distro.
> Of course you may feel that Google Trends is not the best way to measure popularity, but
Well the usual way is measuring what people buy, because then each voter has to do a significant action (and everyone participates equally) of course that doesn't work very well for Ubuntu ... at least atm.
Vignatti: X Census (for 1.9) Posted Sep 6, 2010 23:32 UTC (Mon) by dlang (✭ supporter ✭, #313) [Link]
as someone who has been using linux since the SLS days, I can say that the reason I have the people I support running ubuntu has nothing to do with the free CD program and everything to do with the fact that they very quickly jumped into the lead in terms of making a distro that would 'just work' without me needing to tweak things for the particular hardware.
Since then the other distros have improved in this area, but given the rate of progress prior to ubuntu, I definantly credit them with spurring progress in this area.
Vignatti: X Census (for 1.9) Posted Sep 3, 2010 4:05 UTC (Fri) by dbnichol (subscriber, #39622) [Link]
To be fair, a guy named Chase Douglas from Canonical has been a major force in the multitouch work that will go into the next server release. I don't recall any Canonical or Ubuntu folks around the X lists prior to that. Here's an example of very nice upstream involvement:
http://lists.x.org/archives/xorg-devel/2010-August/012036...
Vignatti: X Census (for 1.9) Posted Sep 3, 2010 4:56 UTC (Fri) by Milan (guest, #26716) [Link]
But his work will be completly Canonic centric again. Their will not be able to merge it upstream as their server centric model has been broken by design from the beginning. Because they ask nobody and made it in house as other. Too sad, Canonical.
Vignatti: X Census (for 1.9) Posted Sep 3, 2010 12:50 UTC (Fri) by dbnichol (subscriber, #39622) [Link]
This is in no way the first time someone has implemented an extension and put it in their distribution to test before it came to Xorg. Do you think DRI2 would be a reality if Fedora hadn't pushed a whole not-upstream DRI/DRM stack on their users?
It's an extension; that's the beauty of the process. If the extension doesn't get approved or needs to be altered, nothing changes for anyone else.
Vignatti: X Census (for 1.9) Posted Sep 3, 2010 12:53 UTC (Fri) by alankila (subscriber, #47141) [Link]
Damned if you do, damned if you don't...
Vignatti: X Census (for 1.9) Posted Sep 5, 2010 21:48 UTC (Sun) by jospoortvliet (subscriber, #33164) [Link]
No, only damned if you do it wrong. To quote the parent:
"Because they ask nobody and made it in house as other."
They could of course ship it with Ubuntu first, if they wanted. They should however have discussed the approach they choose with the *experts* before starting to write it - so they knew in advance if it was the right one. Now they wrote some useless code - because it is the wrong approach.
Google made the same mistake with a lot of Android kernel code - and got exactly the same response. But you can do things different. It's not even rocket science if clearly both small (Collabora, KDAB, KO) and large (Nokia, Intel, Novell) companies can do it right.
Vignatti: X Census (for 1.9) Posted Sep 3, 2010 17:10 UTC (Fri) by whitemice (guest, #3748) [Link]
>It's not just Jef. Canonical's approach to this issue has been utterly
>shambolic and they thoroughly deserve most of the flak they get.
+1
It wouldn't be an issue at all - Ubuntu people just need to stop talking like they *created* the software. That is what grates people.
Vignatti: X Census (for 1.9) Posted Sep 2, 2010 16:08 UTC (Thu) by tzafrir (subscriber, #11501) [Link]
OTOH, I was not aware of the role of Oracle in X.
Vignatti: X Census (for 1.9) Posted Sep 2, 2010 16:18 UTC (Thu) by rahulsundaram (subscriber, #21946) [Link]
It is actually the role of Sun Microsystems, now being attributed to Oracle after the acquisition. Helps to keep that in mind.
Vignatti: X Census (for 1.9) Posted Sep 2, 2010 17:13 UTC (Thu) by rahulsundaram (subscriber, #21946) [Link]
I am not surprised with that. However Novell seem particularly low compared to kernel or Xorg contributors. Anyone know why this is the case? I am genuinely curious.
Vignatti: X Census (for 1.9) Posted Sep 3, 2010 3:58 UTC (Fri) by dbnichol (subscriber, #39622) [Link]
If you include SUSE Germany, Novell actually has a small group of guys that have been working X for a long time (Matthias Hopf and Stefan Dirsch come to mind). They haven't done a lot of core development lately (besides Matthias doing panning in Randr), but they usually at least contribute some good bug fixes. Egbert Eich did a ton of X development for SUSE until the last few years.
Vignatti: X Census (for 1.9) Posted Sep 2, 2010 18:12 UTC (Thu) by armijn (subscriber, #3653) [Link]
Yeah, let's forget about things like sponsoring of free software events, like the party at GUADEC (parties are just for social cohesion and face time, but don't give you code after all), the shipping of free CDs/DVDs worldwide to install parties and so on. Just code, that's what counts.
PS: I don't code
Vignatti: X Census (for 1.9) Posted Sep 2, 2010 18:43 UTC (Thu) by Los__D (subscriber, #15263) [Link]
I am SO tired of this whining about Canonical.
Does it really come as a huge surprise that a company with a revenue of 30 million and 350 employees contributes much less than i.e. Red Hat, with a revenue of 750 million and 3200 employees?
Vignatti: X Census (for 1.9) Posted Sep 2, 2010 19:02 UTC (Thu) by klbrun (subscriber, #45083) [Link]
When you compare annual revenue to number of employees (a handy benchmark), it becomes clear that revenues are probably not the only support for Canonical's employees. Red Hat, on the other hand, looks about right for a software company.
Vignatti: X Census (for 1.9) Posted Sep 2, 2010 19:33 UTC (Thu) by Los__D (subscriber, #15263) [Link]
Maybe, but that isn't really relevant to my bitching. :)
Vignatti: X Census (for 1.9) Posted Sep 2, 2010 19:55 UTC (Thu) by farnz (guest, #17727) [Link] No, but it comes as a surprise that my employer (1/10th Canonical's size) contributed at least as much useful code to X.org as Canonical did, given that X.org is very visible to Canonical's users.
Vignatti: X Census (for 1.9) Posted Sep 3, 2010 2:25 UTC (Fri) by bk (guest, #25617) [Link]
I wasn't aware that Canonical owed you anything. Perhaps X works well enough for them as-is, or the decision makers simply decided not to donate paid company resources to X development. I don't believe there is anything in the X11 license precluding this.
Vignatti: X Census (for 1.9) Posted Sep 5, 2010 0:47 UTC (Sun) by briangmaddox (subscriber, #39279) [Link]
Turning everything on here (and other websites) into a pi!@#$% contest about contributions is surely a good way to convince people about the merits of the FLOSS community and come over to our point it view and enjoy the same freedoms we do.
We get it kids, now let it go for a while. These flame fests aren't going to convince Joe Schmoe to try out FLOSS, and they certainly aren't going to convince Joe CTO to adopt FLOSS and start contributing back to the community. Face it, if we look like a bunch of people who can't get along with each other, we're not going to convince others we can get along with them.
We know Canonical doesn't contribute much upstream. We get it now, really. Let's do something to enhance the signal to noise ration on here instead of degrade it :)
|
|
||||||||||||