mjg59: This thread on the Sun initiative to pro

Matthew Garrett (

mjg59) wrote,
@ 2007-06-01 02:22:00

Entry tags:

advogato, ubuntu

This thread on the Sun initiative to produce a more Linux-like distribution of OpenSolaris is interesting not so much for the technical goals, but as an insight into the community. The immediate response to the proposal? A set of complaints that it hasn't followed the correct process.

Now, to be fair, there have been other threads discussing the technical and social desirability of such a distribution already and some are still ongoing - it's not like the OpenSolaris community has ignored it entirely in favour of process fetishising. However, it's indicative of a community that's more concerned with procedure than it is with producing high quality software. What I suspect has happened is that Sun felt it was important to ensure that procedures were put in place before creating a community in order to avoid the problem of it immediately turning into low-grade anarchy. The problem with this is that if you create procedures before you create community, the people who end up enforcing the procedures tend to be the sort of people who find enforcing procedures to be the interesting part of the job rather than the ones who see them as necessary evils to enforce moderately sensible community development.

Sun's clearly stuck between a rock and a hard place here. They don't want to be seen to be directing the community too strongly, or they'd end up like Red Hat did with Fedora - you alienate your community by giving them the impression that they have no influence over the direction of the distribution. On the other hand, the rapid growth of the Ubuntu community strongly suggests that having a strong leader can ensure that people don't get too bogged down in procedure at the cost of technical progress.

I suspect that Sun made the wrong call here. Having more corporate influence to begin with might have put off some contributors, but it would also have allowed the development of a functional community at an earlier stage. Sun could then gradually withdraw (as Red Hat have been doing with Fedora) and let the community play a larger role. Perhaps that way we'd hear more about the technical contributions coming out of OpenSolaris and they wouldn't be left with just the same two cards they've been playing for a couple of years now - dtrace and ZFS.

On something of a tangent, I don't think their choice of license has helped a great deal here. Two of the areas where OpenSolaris is most behind Linux are hardware support and desktop/kernel integration. The choice of a GPL-incompatible license strongly discourages any Linux developers from working with the low-level code of OpenSolaris due to the perceived risk of contaminating the Linux codebase at some later point. That rules out a huge pool of skilled developers (I'm going to be potentially controversial here and say that I suspect Linux has a larger skillset here than all the BSDs put together, and quite possibly Solaris added to that) who are writing a good deal of advanced code. Meanwhile we're left with OpenSolaris having even worse laptop support than FreeBSD. These are people who could have added to the community socially as well as technically. Instead you've got a community led by a set of Sun staff who are paranoid about doing anything that might make it look like Sun controls the community. Score.

(Post a new comment)

(Anonymous)
2007-06-01 07:10 am UTC (link)

Not so much "contaminating the Linux codebase" as "contributing to something walled-off and thus not useful". A GPLed OpenSolaris would likely have resulted in some shared code between Linux and OpenSolaris, and despite common perception, I doubt the sharing would go in only one direction. A CDDLed OpenSolaris means reimplementing things from one kernel in the other, which gives me a strong feeling of "Yes, but Linux works fine, so why bother reimplementing it?".

(Reply to this)

	zaitcev 2007-06-01 07:23 am UTC (link)
	I thought that alienating the ballast members who remove libcomerr2 was a good thing. But seriously, I know several people who were alienated by Debian and escaped to Gentoo and Fedora, simply because they would take patches and Debian won't. (Reply to this)(Thread)

	broonie 2007-06-01 08:57 am UTC (link)
	Another complaint that gets heard about Debian is that it takes too many patches :) (Reply to this)(Parent)

	(Anonymous) 2007-06-01 10:53 am UTC (link)
	Maybe that is why Debian has aptitude and Fedora now is stuck with yum... there needs to be some thought and some filtering, at some level. Just don't make it central. (Reply to this)(Parent)

	pjc50 2007-06-01 01:03 pm UTC (link)
	Why does anyone need opensolaris? What does it offer that you can't do on one of the Free unix clones? (Reply to this)(Thread)

	mjg59 2007-06-01 01:08 pm UTC (link)
	zfs and dtrace. Opensolaris - not just a one trick pony! (Reply to this)(Parent)

	jmtd 2007-06-01 03:27 pm UTC (link)
	Zones; the Service Management Framework too. (Reply to this)(Parent)

trs80 [typekey.com]
2007-06-06 12:48 pm UTC (link)

Also integrated Fault Management, and an NFS server that performs consistently between releases (unlike, say, Linux's http://linux-nfs.org/pipermail/nfsv4/2007-January/005562.html "OK, but I have no idea why that's happening." (performance increase in 2.6.20, after a 80% reduction 2.6.15 to 2.6.17).

Back to the community, I think Sun having more corporate influence wouldn't matter, since the culture of procedure is very prevalent within the company, and probably what caused the emphasis in OpenSolaris in the first place.

(Reply to this)(Parent)

ajaxxx
2007-06-01 01:26 pm UTC (link)

This is only slightly related, but it's not completely accurate to describe Fedora's slow opening as a conscious strategy on Red Hat's part. We've been trying to get a unified and open Fedora since, well, RHL9. It's the execution that's sucked.

I'm not convinced it's a strategy worth emulating. It definitely gets you a large set of core contributors - namely, your employees - but interacting with the masses is hard and you have to show that you trust them and are willing to work with them and man is that tough to do when you're serving two masters. I should have put a lot more time into F7, but instead it went into RHEL5. We have a unified Fedora now, but we still haven't really figured out how to staff engineering so that we have serious full-time Fedora work. Therefore, no matter how much time we put into Fedora, we're always going to take abuse for Fedora being a RHEL beta.

Whereas, if you're Ubuntu, you only have one OS, so there's no question of resource allocation.

Of course, now I'll let my bias show and state that I think the Fedora/RHEL model is a better way to build a product you can actually sell. Dapper is sort of a tacit admission of this. But it's a brutally difficult trick to pull off, and we're still screwing it up on a daily basis. Sun's got their work cut out for them.

(Reply to this)(Thread)

(Anonymous)
2007-06-01 09:00 pm UTC (link)

I'm a bit curious what you mean by that Dapper comment. I was under the impression that RHEL was more than a normal release with a longer support cycle, but in truth I'm not all that familiar with the for-pay distros. But I think aggressive (http://www.pnaelv.org/letter_body/) defense of trademarks against (ironic how defending one's trademarks can be an offensive tactic) people using source suggests that some within agree that it's the source code and not the organization that makes the distro valuable. Hell, resource allocation doesn't make any sense if RHEL and FC aren't substantially different things.

Where I think Ubuntu went right is that it was founded by people who knew fully well how the community system worked, people who knew what was wrong with the status quo and set about proving their theories right. Thats why the Code of Conduct was presented early on, why the community was engaged in how the process would work, and MOTU was created. The education effort on "community" felt hokey to me, but coming from a Debian background I guess I understood this better than others.

(Reply to this)(Parent)(Thread)

ajaxxx
2007-06-02 01:42 am UTC (link)

Now I'm the one that's confused. Red Hat told CentOS to stop calling it a Red Hat linux because that's the name of the company, and if you don't defend the trademark then you lose it. The goal there is to have a controlled message about what "Red Hat" is as a brand. We're more than happy to have people using CentOS; it means if they ever decide they want paid support then RHEL is the logical choice. But calling it Red Hat is one step away from being kleenex or coke.

Canonical take a different tack here, but part of that is that their OS doesn't share a name with their company. You can be assured that if it were Canonical Linux, the KDE variant would not be called Kanonical. They set out from the beginning to build a brand that was not directly connected with the company behind it. Red Hat, on the other hand, was a distro before it was a commercial product, and had tremendous cachet in those two words as the name of the distro just because it was "the Linux you can buy". But RHL made very little money, on pathetically thin margins. When we switched business models, calling the new product anything besides Red Hat Something would have been suicide, because, uh, the bubble was popping. Lesson in brand building, I suppose.

All of which is aside. My point with about Dapper is just that you can't actually sell a distro with a six-month release cycle and one year support span, with massive upheaval between releases. No one's going to buy Fedora 7 unless they know that it'll be supported in three years, any more than they'll buy Gutsy. If you want to sell an operating system, you have to have some engineering facing forward at new development, and some facing backward at the shipping product. The only real distinctions between Dapper and RHEL are those of scale and clientele.

The weird thing here is how people talk about Ubuntu as being disjunct from Canonical, but not Fedora being disjunct from Red Hat. If Canonical were to fold tomorrow, the only reason Ubuntu would stand a chance of continuing on as a viable distro is because it's populated by Debian people who are sick of Debian. Yet, Fedora is derided for being Red Hat dominated. Very interesting bit of social engineering there.

I completely agree that Canonical has done a better job of community building than we have. But the primary difference seems to be one of messaging and execution, not actual strategy.

(Reply to this)(Parent)(Thread)

(Anonymous)
2007-06-03 02:27 am UTC (link)

I'm not very clear on the history of Red Hat, Fedora, CentOS etc; I can't really say whether Fedora is dominated by RH or not (or if it really matters), or what their community building strategy was. But looking over the Centos website in archive.org, I don't see anything extremely inappropriate with their references to Red Hat on their site. I certainly don't think they were representing themselves as Red Hat, or making substantial changes and calling it RHEL.

If Red Hat felt exactly the same as you wrote, I think they would have gone about the CentOS thing differently, but I'm sure when CentOS saw the cease and desist they responded in typical Open Source Project fashion: refuse to come to the negotiating table, followed the request to the T then made snarky comments about it all on Internet venues. To this day they don't link to or regularly describe themselves as Red Hat based, even though they cater very specifically to people who want RHEL without pricey RHEL support. Amusingly, Google Adsense still picks up on the association, and Red Hat's losing some of those converts you hypothesized.

Ubuntu has a name too. And a trademark. And it happens to be owned by Canonical. They're different though, so I guess, lesson learned :P Personally, I don't think that Ubuntu is separated enough from Canonical, though the trademarks probably need to be held by an entity like Canonical, the same as the SPI holds Debian's marks. Looking at the release goals, it's clear that Canonical's domination has some weird side effects. Theres a huge number of maemo-mobile related goals that I believe Canonical is being paid for. And I'm not real clear on what "support" within Ubuntu really means. Security patches mostly, from what I've gathered. You're right that Ubuntu survives on the strength of disgruntled Debian users and developersif Canonical fold (and the ten million held by the Ubuntu Foundation), but I don't see why that's a problem or how it's related to Fedora being RH dominated.

But thanks for clarifying the Dapper thing. That makes sense, though I don't think anyone really thought it was wrong. There's clearly server people who don't feel that 18 months (not a year ;) is proper and need secured environments. I suspect for the right price you could buy LTS for anything though. And I'm told that people are buying Feisty through Dell, so maybe you can sell a distro on six month release cycles ;)

But certainly, Sun's failed to quickly address community orientation and move on to bigger and more interesting problems, of a more technical nature.

(Reply to this)(Parent)

	(Anonymous) 2007-06-04 04:25 pm UTC (link)
	There's one huge problem with your assumption. You are assuming that there was no community, which is wrong. The OpenSolaris community is almost two years old now. There was a community before the proposal was made, hence the expectation to follow the procedures. (Reply to this)(Thread)

	mjg59 2007-06-04 04:59 pm UTC (link)
	What? No, I'm not making that assumption at all. (Reply to this)(Parent)(Thread)

(Anonymous)
2007-06-05 02:13 am UTC (link)

Well, it sure seems like you are.

Sorry, but I don't agree with the whole "throw the procedures out the window" thing.

I have been involved in far too many projects that just started without any procedure at all, and then ended up blowing up when procedures were attempted to be added because "we never needed them before!"

Besides, remember that one of the original guiders of this community is Roy Fielding, who helped found the Apache Foundation.

These are not people without experience in doing these things.

Not only that, there was technically a Solaris community and even a pilot program before there were ever procedures in place.

In fact, all of these procedures were technically only ratified very recently (last couple of months).

So, if you want to be all grumpy about it, we had a community, and almost two years later, we have procedures.

(Reply to this)(Parent)(Thread)

mjg59
2007-06-05 02:24 am UTC (link)

No. The point isn't to have a community without procedures. The point is to have a community that acts as a community, not a community that acts as a bureaucracy. You need to prove you can act as a community before the procedures exist, and then once those procedures do exist you need to (and this is the important bit) continue pretty much as you did before. The point of having a community is for people to agree with each other. The point of the procedures is to make it clearer that people agree with each other as the community grows larger. They're there to be enacted when needed, not there to be used as tools to shut down a discussion before it's even started properly.

As an example: I'm going to ignore your comment because you've failed to comply with the procedure for using paragraphs correctly. Please reapply at some later date with your paragraph usage corrected and we'll then be able to discuss the issue.

(Reply to this)(Parent)(Thread)

Good/bad introductions to the OpenSolaris Community personality
(Anonymous)
2007-06-05 02:54 pm UTC (link)

Hi Matthew,

In my opinion, if one wants to judge the OpenSolaris Community, opensolaris-discuss should get a weighting of about 0.1%, you gave it about 99.9%

I admit it's tempting to characterize the OpenSolaris community the way you do here if you only look at the opensolaris-discuss forum.

But look anywhere else on opensolaris.org (/anywhere/), you'll find a huge, vibrant healthy community, "acting like a community". *Anywhere*. Any mailing list, Any OpenSolaris Project Group, any OpenSolaris Community Group. Everywhere you'll see overwhelming flexibility, diversity, non-zealotry, and non-fear. Except opensolaris-discuss.

It's really, really key to note that, when it comes to Contributors and Core Contributors, opensolaris-discuss is ignored by many of them, maybe even most of them, and in fact, to my knowledge, probably the /vast majority/ of them. Unfortunately, that unmitigated, disasterous melee that many people mistakenly believe was a community discussion about Indiana, took place on opensolaris-discuss.

Eric

--
Eric Boutilier
OpenSolaris
Blog: http://blogs.sun.com/eric_boutilier
Email: eric dot boutilier at sun dot com

(Reply to this)(Parent)(Thread)

Re: Good/bad introductions to the OpenSolaris Community personality
(Anonymous)
2007-06-06 08:07 pm UTC (link)

Which raises an interesting point: both Ubuntu and Open Solaris (I'll take your word on this) have a problem with segregation via forums. It's an unfortunate but true statement that few people participate in both forums and mailing lists. Allowing this segregation essentially means allowing the community to split into factions of developers and users, often with very poor communication channels between them.

Do portals that integrate mailing lists and web forums exist? And if they did, would either Ubuntu or Open Solaris accept them? In both cases I think you'd have trouble deciding a) how to merge (or keep seperate) the histories of both, and b) how to resolve the threading differences between them. Oh and I guess c) whether it's worth the effort.

(Reply to this)(Parent)(Thread)

Re: Good/bad introductions to the OpenSolaris Community personality

mjg59
2007-06-06 08:10 pm UTC (link)

The Ubuntu mailing lists were gatewayed to the forums at one point, but it led to problems with the development list so that (at least) was disabled. More recently we've established Ubuntu Forum Ambassadors who have responsibility for bringing issues raised on the forums directly to the development community. It's a bit too early to see how that will work out, but I think the early signs are promising.

(Reply to this)(Parent)

(Anonymous)
2007-06-04 08:08 pm UTC (link)

I think his point is that the community is dysfunctional. Some are focused too much on procedure, to the point that dedicated people like Ian Murdock and friends have trouble apparently accomplishing their goals within Open Solaris' umbrella. And nobody it seems notices that a wiki would be a more appropriate venue for suggesting changes to the proposal, etc.

And note well the lack of appreciation for the proposal as well. I don't bother following Open Solaris's efforts, but they must feel that source based operating systems inevitably lead to multiple incompatible binary distributions, so they want to support that rather than fight that. Which is why I just ordered two DVDs from them that contain 4 distributions. That's not a clusterfuck waiting to happen or anything. Hopefully I'll just let the kit gather dust rather than try to figure out which one I want to try. Because it seems hard to get excited about what can be accomplished when nobody's sure they want to see it accomplished at all. And that's the most dysfunctional part of the whole thing.

(Reply to this)(Parent)(Thread)

	(Anonymous) 2007-06-05 02:15 am UTC (link)
	No, the point is every community has dysfunctional people. There is plenty getting done. This project, in particular, was poorly proposed and lacks focus. It also seems to lack purpose. Many of the things this project is supposedly asking for are already being done. It was also presented in a rather poor fashion initially. First impressions do matter. (Reply to this)(Parent)

	The easy fix ceri.myopenid.com 2007-06-05 12:31 pm UTC (link)
	Instead of complaining that the OGB are being obstructive, just follow the process. Executives the world over attempt to railroad due process and I think that allowing this would really, really piss off the community especially when all they have to do is find a group to nominate it. (Reply to this)(Thread)

	Re: The easy fix (Anonymous) 2007-06-06 08:18 pm UTC (link)
	There's a world of difference between doing and voting. Some people say "vote with your feet", but I think the best approach with an open source project community is to vote with your hands. You shouldn't need "community approval" to do things. Maybe Project Indiana is different because it wants to change the community, I'm not sure. (Reply to this)(Parent)

Process versus community.... and progress

http://openid.sun.com/comay
2007-06-07 11:55 pm UTC (link)

Matthew, I think your characterization of the community being more concerned with procedure than producing high quality software is unfair.

First a disclaimer and point of introduction - I work at Sun in the OpenSolaris virtualization group. Although the past few years I worked on and lead the Solaris Containers (zones) project, for awhile now I and a number of other engineers have been looking at making OpenSolaris more "familiar" (to use Ian's word) to users not familar with the OS. This work includes looking at the way Solaris is packaged (or its product structure) to what tools chains are included to how we can grow the OpenSolaris community itself and include a much greater set of maintainers who are not employed by Sun.

The OpenSolaris community is relatively young - it will soon celebrate its two year birthday although obviously the related and intersecting Solaris community is much, much older. In fact, it was only recently that the OpenSolaris constitution was ratified and put into effect. As such, I think there were some very legitimate concerns around some aspects of the proposal that was presented around the areas of branding and goverance. So although many of the goals of the Indiana project align nicely with the work that we're doing, I think the "process" here helped uncover some of those concerns I spoke of earlier. I think that's healthy actually and I suspect the project in question will end up stronger as a result.

As for producing high quality software being of lower priority, I'd like to second someone else's post about checking many of the other forums or mailing lists. There is a lot of active technical discussion going on a wide variety of topics from existing technologies (Zones, ZFS, SMF) to new work being done (new Caiman installer, enhanced power management) to other areas like documentation and the OpenSolaris website itself. And though the vast majority of the code changes are being done by Sun engineers, more and more of the engineering work (from initial discussions to design and architectural reviews to code reviews) are being done outside. As the community becomes more familiar with the code and the technology, I fully expect the code changes from outside of Sun to pick up (we still have some infrastructure things we need to finish as well.)

There are a ton of changes going into OpenSolaris - heck if you want some weekend reading, there's a hundred or so pages in http://docs-pdf.sun.com/820-0724/820-0724.pdf that documents much of the changes since Solaris 10 was shipped. In the past month, a fairly complete set of wireless drivers were putback into OpenSolaris along with a new network configuration service which enables sensible, auto-configuration of network ports on machines such as laptops. So while there is indeed a lot of heat and not much light on general mailing lists like opensolaris-discuss, I think the community is far more united and working effectively than that mailing list would indicate.

I do appreciate your observations though about the rock and the hard place. It's indeed an interesting situation and one that I'm confident our young community will actually come out stronger because of it.

Finally, I respect your opinion regarding the license although I disagree with respect to its ramification. First of all, I don't think many of the Linux device drivers would be easily portable to OpenSolaris even if the licenses were compatible. And many of them would need additional work in order to support some of standard OpenSolaris features such as FMA, the GLDv3 network device framework, etc. That's not to say that we don't need more drivers - we do although again, there have been great strides in the past few months with the wireless drivers, Xorg 7.2, additional SATA support, USB videocams, etc.

(Reply to this)

My blog about pharma care
(Anonymous)
2008-01-25 08:05 pm UTC (link)

Why you should buy buy viagr?
First of all, if you buy us viagra alternatve you get quality service and great discounts!
Why would you need buy iagra?
Viagra in the first place make a uverenost, most of the problems are solved in the privacy of using our products. We will help you learn infinite possibilities in the intimate lives.
Why do you have to use our shop?
You pay only real value of the goods, without deductions resellers percent! Ordering cheap viagr us, you get free delivery and discounts on other products.
On our site you can order viaga online. The lowest price. Large discounts. Full anonymity. Free shipping.
More info about evitra you can be found on our website http://buy-viagra-online-pharmacy-in-canada.info/buy-cheap-viagra-online-now/alternative-female-viagra-viagra.php

(Reply to this)

	order xanax online (Anonymous) 2008-02-04 03:34 pm UTC (link)
	Different kinds of xanax. Influence of xanax on your health. picture of xanax pill alprazolam buy xanax buy xanax without prescription in usa xanax and serapax xanax online side effects of xanax buy xanax without a prescription (Reply to this)

	http://www.inpaz.com (Anonymous) 2008-04-11 04:50 am UTC (link)
	Just test (Reply to this)

Linux and Solaris

rockrags_us
2008-05-22 06:24 pm UTC (link)

Hi everyone,

OpenSolaris scales on any hardware, has built in virtualization, great web service infrastucture, fault management, diagnosability, and much more. OpenSource tactics will be implemented in the Future business world as well. And I thing that Linux is a next century role model.
Linux and Solaris should really see how much they can work together. Each has some really great features and co-operation here can only be beneficial.

rockrags.us

(Reply to this)

	TigssAicIs (Anonymous) 2008-05-22 09:40 pm UTC (link)
	vg8Pmn jr39ug7djalfgpitg94gbvm (Reply to this)

Mini Sitemap:

Username:		Create an Account Forgot your login? Login w/ OpenID
Password:
	Remember Me
	English • Español • Deutsch • Русский…