A tale of two cultures...
First, here's a quick status update on the Gentoo Offer - Grant is the remaining trustee - it appears that everyone else has resigned - and Grant was away for the weekend. He is now apparently back in town, but as of today (Tuesday, Jan 15th) I have not heard from him yet. UPDATE: Just got an IRC message from Grant - he is currently trying to get caught up and asks everyone to subscribe and post to gentoo-nfp@gentoo.org rather than the trustees alias. (Send email to gentoo-nfp+subscribe@gentoo.org to subscribe to the list before posting.)
Now, back to my main blog post.
At this point, it's pretty clear to me that there is a huge disconnect between Gentoo developers and the rest of the universe - by "universe" I mean Gentoo users, outside Gentoo-based projects, organizations using Gentoo Linux and the Open Source/Free Software community as a whole.
If you ask a typical Gentoo developer and a typical Gentoo user to explain the reason for this disconnect, you may find yourself getting two completely different answers.
The larger Gentoo community tends to view the Gentoo project as a cathedral of sorts. There is a priesthood of developers who control the direction of the project. There is very little communication coming out from the cathedral to inform users about what is going on. Changes in the Portage tree often seem to come with little or no warning. The Foundation has essentially no voice, the Council only communicates with fellow developers, and Gentoo users are left to fend for themselves on the forums, gentoo-wiki.com and sometimes in a bugs.gentoo.org bug report.
Of course, it is always possible for a user to become a developer. If a user would really like to have insight into ways of the Gentoo priesthood and have the possibility of making a difference, they must go through a long application process and become experts in Gentoo's internal policies and procedures. After which, they must subscribe and participate in a large number of mailing list threads in order to get a basic understanding of the project as a whole. This takes hours and hours of time, and will require many additional hours every month. Eventually, the developer may find that they are able to successfully contribute to a specific sub-area of the project and begin to make a difference. They have now become a Gentoo developer.
I should point out that very few Gentoo users, even the extremely skilled users who use Gentoo in various innovative ways either personally or in a professional capacity, can afford to make the necessary time commitment to join the elite priesthood of Gentoo developers. This is unfortunate.
Now, let us enter the cathedral of Gentoo developers. In it, you will find many developers quietly working away on various projects, and overall making decent progress and doing good work in their particular areas of focus. Yes, many developers are aware that some things aren't getting done - such as Gentoo Weekly News, new Gentoo releases, addition of packages to Portage, and moving of certain packages to stable. But overall, most of the time, things are moving forward at a decent pace. Users may not get to see all the fruits of the Gentoo developers' collective work, nor be very well-informed of recent changes and improvements, but that doesn't mean that things aren't happening.
Gentoo developers are aware that many users are upset that certain things aren't getting done. From the perspective of a typical Gentoo developer, the solution to this problem is for more users to become developers. Those who complain about lack of progress without volunteering to help solve the problem are simply viewed as whiners who are more than happy to complain but unwilling to help.
Meanwhile, Gentoo's user community continues to get more agitated even as Gentoo development continues - since it is not happening at a pace that Gentoo users generally find satisfactory or acceptable.
So, what can be done? Here are some ideas.
1) First, generally speaking, communication and transparency from the Gentoo project is poor. The Gentoo Foundation needs to ensure that the lines of Gentoo communication are open and that users are informed of new developer innovations and that developers are informed of user needs and perspectives. The Foundation needs to focus on building bridges of communication between the various groups of the Gentoo community.
2) There are many capable, qualified, motivated Gentoo users who are willing to help Gentoo in some capacity but do not have a means to do so currently. In general, the current time commitments for Gentoo developership are too high. The Gentoo Foundation needs to look into ways to optimize the developer experience for the benefit of the project as a whole as well as current developers, by reducing the overall time commitment required to be an active, informed developer and maximizing the positive impact that an individual developer can have in a given amount of time.
3) The Gentoo community has hit a several issues related to scale which have essentially placed an upper ceiling on the amount of progress that can be made by the Gentoo project as a whole (I would even go so far to say that this ceiling will exist regardless of the number of developers currently active on the project.) The Gentoo Foundation needs to resolve these scalability issues for the long-term health of the project.
Fundamentally, I don't think the solution is to simply add more developers. The solution is to focus on refining the developer experience and maximizing the impact that an individual developer can have in a given amount of volunteered time per week. If we are not sensitive to this issue then we will not get the quality or quantity of developers that we need to move this project forward over the long term, and an upper ceiling on our progress will continue to limit Gentoo.
There are several approaches for improving the situation - some briefly touched on in this blog post, and others that I'll talk about in future posts.
53 comments:
Nicely said, I agree with you :)
As I'm sure many others will too
To be specific, I requested that those who want their comments to be public should use gentoo-nfp@gentoo.org. If you just want me to read your comment, trustees@gentoo.org is fine, but I'm the only person currently receiving those e-mails.
I think you sum up everything pretty nice Daniel.
Sadly devs have stopped to listen users long time ago. I hope they start to hear even the last minute :<
One sad side of the developer-user disconnect involves developers mistreating users that offer feedback. I've observed a couple of occasions where high profile projects (such as the installer) had criticial bugs dismissed improperly and uncivilly.
It demonstrates the cathedral you describe: the user feels completely unwelcome within the developer community and has no opportunity to criticize the actions of a developer except with the developer himself.
This doesn't tend to promote a healthy project.
May I propose a solution to the scalability issues, and the user-developer distance problem?
Gentoo and the Portage Tree is like CVS and SVN, they are centralized content management systems made of code and people.
I believe we should go distributed, like Git, Bazaar, Mercurial, etc.
Every part of the tree would go to overlays, and every overlay would be treated like first class citizens.
KDE for example, would be an overlay, and even KDE extragears could be an overlay of KDE.
There could also be competiting overlays, why stop them? Two KDEs? Yes, the Free Software community usually tackles the same problem from multiple fronts, how many distros do you know, how many inter-process communications systems are there? And they often merge back to an unified and larger than the sum of its parts solution, DBus is that. Just don't cross the same namespace and you'll be fine.
Gentoo-Core would just be an [obligatory?] overlay of the main tools needed to build this all. Use GWN to tell "Danger, Will Robinson! Expat 3.0 is coming in a month!"
b.g.o would take everyone's bugs, commit access? Just like Git. You trust people, you merge their changes, you'd sync to many overlays, not just one. Documentation? Make Gentoo-Wiki a first class citizen too, port the documentation over there. The same with many Gentoo utility sites. f.g.o's Unsupported Software forum would thrive.
No elitism, bring everyone to the same level.
Are we a meta-distro? Yes, not only that everyone builds their own distro, but that we make Gentoo and Gentoo make us back.
I sent this to trustees:
Hello,
Thank you for taking the time to read this. I've been a Gentoo user since the 1.2 days. I've used it continuously at home since then, and on and off at work. It's sometimes hard to convince managers, IT types, etc. to let me install it.
I'm a developer by trade, and have recently and in the past considered trying to become a Gentoo dev. Every time I've decided against it for reasons remarkably similar to the ones Robbins lists in his latest blog post. It seems it would take a long time to figure out even how one would go about figuring out what to do.
I believe Daniel Robbins is willing and able to improve the current state of Gentoo. He's shown that the has the technical ability by coming up with the project in the first place, and he's shown he's passionate about it by making this latest offer.
I've used my real name and the email I use in bugzilla on this email, but I would prefer to remain anonymous. There are heated opinions on both sides of this issue, and I'd rather stay out of any battles that may be under way.
Thanks,
Me.
I do agree with you on the problem of it being very hard for interested users to participate, that is something that absolutely needs tackling.
I also agree that the communication of the developer community to the outside needs to improve.
I don't completely agree on your suggested solution but we do see somewhat similar problems.
Gentoo is not as bad as people sometimes say technologically but that is often hardly communicated to the users. If someone does not actively look around, reads planet gentoo and the (heaven forbid! mailing-lists) he/she has not really a chance to know what's going on and what's coming up in a month or so.
Right now there are a few developers writing to their blogs quite regularly to keep users up-to-date but that number is too small.
Many other distributions ain't better technologically but when it comes to their user relations. That is (in my opinion) the main issue.
People installing Gentoo should not only be send to the forums but to a page that does not yet exist that outlines the different ways to get involved (mailing-lists, forums, planet larry, planet gentoo &ct). Users need to know how to keep in contact with the project itself and that is a lot of work but it is a lot of work that people less tech-savvy can help with which offers new opportunities for people to help the project.
A while ago I had for example thought about a "gentoo podcast", just like the Newsletter, but in a more modern form that people can throw on their ipods and listen to while driving to work or whatever. Help people to keep up, keep them interested and curious. I think that would help a lot.
Personally, I think there is another side of the coin as well - users seem to think we developers do nothing at all - in that, all we do is work on Gentoo - we don't actually use it (how that works I am not sure - but it sure feels that way.)
zeroum: We (well, some of us) would *love* to move to Git, and we are watching Git development closely, however it is still lacking a few features that we really need. Things like, only portions of history, or only checking out portions of a repository rather than cloning the entire thing. As for things being different overlays - in theory that works great, however, unfortunately, you get applications that depend on this, or depend on that for random (sometimes optional, usually not) features, which if you move everything into its own overlay, you start to get a huge spaghetti string mess of "overlays" - its not that we don't want to do these things, it is that they are unfeasible currently. Hope that helped spread some light on things :)
Could you please clarify whether or not these comments are truly representative of the feedback from the entire user community? Or have some comments been removed? (and, if so, why?)
thanks.
zeroum,
Seemant Kulleen has previously proposed that idea for the main tree. You can probably dig up the blog posts on Google. One major issue is how to deal with cross-repository dependencies. Someone interested would need to add support to the package manager, figure out where the natural splits are in the global dependency graph, etc.
A user mentioned git as a repository ... that idea is a no no. Git is really slow. It is one of the worst versioning systems out there. Bazaar or Mercurial would be better if you want to go distributed. I think the latest version of bazaar is kick ass. :)
Anyways on to my planned post. I am what you would call an experienced Gentoo user. I have been with Gentoo since it's birth, and pretty much stuck around. I can help out with most issues, because I have had them at one point. I was in line to become a dev, but I just got so pissed at Gentoo for awhile that I had to leave. I still kinda harbor ill feelings at certain items. Hell, I recommend Arch Linux to people. I use Gentoo because I am a code tweaker by heart and I can't live without it. I write my own ebuilds for applications and I can do some advanced things in the ebuilds too, not just the plain unpack, compile, install routine. I get into advanced eclasses now and then and things like that. I love working for myself, if something isn't quite right, "Ah well I will save it for later." It's very relaxing and no users breathing down my neck. I have my own schedule and my own plan, and I am comfortable with that.
What I am getting at is my reason why I think some people don't become devs -- the users. To take to point, the whole KDE 4.0.0 drama that went on for the past 4 days. There were people speculating if it will get in portage or not and they were wondering why it was taking so long and all that. It was annoying to me to see it on the forum. I think the devs just ignored the thread. A little bit of time and the users would have found a post on the devv ML about the eclasses that were submitted for QA. The users do not understand that there is a procedure that must be followed. On top of that, they don't search, they don't look for themselves, and they just assume something and what they assume is totally incorrect, and when they don't get a good response, they believe the incorrect thing even more. I literally had to use large letters and yell at a couple people in that forum to tell them the truth of the situation after it was posted several times in that thread. I feel like a bad person when I do that, but sometimes people just need to be virtually slapped. The devs get so much hassle and gripe from users that it's insane. No person should have to go through what the devs go through. I try to defend the devs when I can on the forums and if I need to kick some butt I do. I appreciate what the devs do because I know what they do is not easy. If other users just calmed down and stop being childish whiney gits, more people might become devs. They know that if they become devs, they will go through what they see on the forum.
The problem is nothing more than stupid users. Seriously. I think I vow from now on that if I see a user being stupid, I am going to tell them off. I am completely pissed at them.
Gentoo is a victim of it's own success. Back when the userbase was smaller, the devs had more time and things got done quickly. They talked on the forums and were great to get along with, but now they need to answer to all the people giving them shit. No wonder they don't talk to users much. This is why they are "the priesthood that you don't see very much locked away in their church".
Sorry for my rant, but it's the truth. I know some people will read this and go, "Is he talking about me?". If you need to ask yourself that, then you are guilty of what I say because you will realize that you have done the things I said.
Cheerio.
One thing I see which Gentoo does is create far too many Gentoo projects which has serious problems with #2 and #3. Any time you add a sub-project to Gentoo it will add more complexity and take more time.
Gentoo Weekly Newsletter, great idea and it did run for quite a while very successfully. However it takes many hours per week to publish it. These valuable hours could be spend on other things such as improved documentation. I wouldn't suggest fully dropping the newsletter but scaling back to monthly has been a good idea.
Gentoo overlays/sunrise/bugzilla. Where is development supposed to be focused? We have a main tree and several other areas where development may be! I don't care if you need to switch revision control system, but there needs to be some simplification on the development process.
Figure out the core parts of Gentoo which are essential to the survival of it and the success of it and improve them.
Thank you for this post Daniel.
I think you are very right in your observation of two cultures, this has been even more evident to me in the last few days where I have been participating a bit in the discussion among other users on primarily The Gentoo Forums.
I think your proposals here make good sense and it is nice to hear more about your ideas for Gentoo.
No matter if your offer is accepted in some form or not your writings have already had a positive effect on the Gentoo community so hopefully something good will come of all of this, things seems like they have started moving for change ... now it is just a question of directing this movement into something positive.
Zeroums suggestions (comment #5 to this post) also seem very sensible to me many of these things are already being implemented to some degree, but it is my feeling that they drown a bit because the developers also sometimes have difficulties crossing this "gap in cultures" that Daniel has so well described in his post. For example I wasn't aware of the overlay "rising sun" until a few days ago which is really a great initiative that I plan into looking a bit more into, it is a great way to do some bridging of the gap, things like this should somehow be better communicated to users like me.
Please keep your posts coming Daniel, I really enjoy reading them :)
The Gentoo forums have a reputation across the Linux world as a place people can come to find solutions to their problems. Left to fend for themselves, evolution has made Gentoo Users adapt and create something that has become valuable to many others.
I'm not knocking anything you said, just pointing out one of the upsides to the current situation.
@Zeroum : personnaly, I'm stronly against Overlays. When I switched to Gentoo more that 2 years ago, I was incredibly pleased not to fight with repository as I've done before with Fedora.
An Fedora, for example, is doing exactly the opposite : they have grouped repositories.
This is just my point of view.
And yes Daniel, as a Gentoo user, I confirm that I feel totally ignored by devs. That doesn't mean they're not doing a great job.
Alexis, a French dev, often comes on the forum to give a hand and communicated a lot when he achieved an amazing job about Texlive.
Hum, according to me, every organisation must have a Leader who takes the decisions.
If it's not the case : voting without flaming!
I believe we should go distributed, like Git, Bazaar, Mercurial, etc.
Well, we definitely should abandon CVS, it's a horrible VCS.
Every part of the tree would go to overlays, and every overlay would be treated like first class citizens.
You can have sort of official overlays which are supported (like those on overlays.gentoo.org), but it's completely unviable for every single overlay out there to be supported by Gentoo. Some of those third-party overlays should preferably cease to exist since they are causing more harm than good.
KDE for example, would be an overlay, and even KDE extragears could be an overlay of KDE. ... Gentoo-Core would just be an [obligatory?] overlay of the main tools needed to build this all.
With current state of EAPI and package manager capabilities, no way to accomplish this short, or even medium term. You'd get a horrible dependency breakage pretty much everywhere, and the amount of work needed achieve this goal is huge. Additionally, huge desktop projects like KDE/Gnome/XFCE etc. are really a bad candidate for this. OTOH, things like games, scientific packages, webapps etc. could work quite well in this way.
There could also be competiting overlays, why stop them? Two KDEs?
See above. This actually breaks things unless the things are compatible, which they by definition will not be because then there would be no reason for such competing overlays to exist.
You need a sane way to express cross-repo dependencies, you need a sane way to manage overlays order/preferences, you need a sane way how to handle overlapping eclasses, etc. etc. etc. Non-trivial at best.
Use GWN to tell "Danger, Will Robinson! Expat 3.0 is coming in a month!"
Does not work. Not because GWN is dead ATM, but because people generally ignore the communication channels. As an example: apache/php transitions/deprecations etc. have been extensively announced on GWN, on forums.gentoo.org, on developers blogs, etc. etc. etc. - yet when the change his the tree, we have been overwhelmed by bugs which for the large parts were not any bugs, but a design change, announced months and months in advance.
Similarly, users ignore existing tools such as elog features and don't read any important information provided by the ebuilds themselve. GLEP42 (Critical News Reporting) tried to tackle this from a different angle, i.e. providing preemptive info as opposed to ex-port info that's offered by elog stuff, but I'm skeptic about what kind of improvement will this actually mean. Simply said, all efforts on communicating packages-related info to users failed to some extent.
Documentation? Make Gentoo-Wiki a first class citizen too, port the documentation over there. The same with many Gentoo utility sites. f.g.o's Unsupported Software forum would thrive.
Pardon me, but hell no. The gentoo-wiki docs contain horrible errors and there's no way to maintain a reasonable documentation standard when anyone can change the provided information at will. Me myself have tried to correct some of the information there once we started to get flooded by 'bugs' caused by poor-quality advice provided on gentoo-wiki.com, only to find out that some 'smart' user broke it the next day because he obviously knows better.
Moderated wiki? Possibly, but we don't have the human resources to provide for such moderation ATM.
Thanks for the update and the explanation, Daniel. I can see the problem of transparancy you speak about now, altough being a gentoo user, i did not have any idea the things were working that way...
I really hope for a happy outcome from this situation, as I am a bit affraid by all the people who announce the death of the distribution.
Thank you Grant too, to be the "last one".
Hi Daniel, I completely agree with you and I hope you lead the Gentoo Foundation again (de facto).
However, as much as I admire you and am thankful to you for creating Gentoo, I think that you should also take a fair part of the blame for the current situation. Simply because you abandoned Gentoo to go to M$. It is your life - but IMHO you should have stayed as a leading dev and president of the Gentoo Foundation.
Still, I hope you lead Gentoo again...
Concerning:
2) There are many capable, qualified, motivated Gentoo users who are willing to help Gentoo in some capacity but do not have a means to do so currently.
Do as Arch Linux do with the AUR (ArchLinux User Repository). This way they have two central places for packages:
* main repository (with Core, Extra, Testing, Unstable sections)
* user repository (with Community and Unsupported sections)
Actually this benefits all devs and users, and is sufficently transparent. Trusted users can move packages to the AUR/community section (possibly adding binary packages), while everyone else can submit package recipies to the AUR/unsupported (no binaries this time). It works great - when no package is in the Core, typically an active user (with lot less participation requirements) already has added the desired package in Community. For example, a Polish translation of the newest OpenOffice.org is always in Community, while only a handful of translations are in Core.
Of course, packages are imported from AUR/community to the Core, but this is done on the popularity basis. AUR-registered users can vote on their preferred packages so after approval they are imported by a developer. Much cleaner solution IMO.
Some of us, developers, are still with you; that is, the ones that resisted and endured the pressure of the priesthood.
Keep it up Daniel, you're our only hope.
As far as I can tell, you really got it.
One of the causes for this is in my opinion, that _everything_ is decided by devs.
I really think, gentoo should have a separate group of people, focusing on PR and keeping the GWN (or GMN) up, releasing new, bugfree releases and such things.
Maybe we should also introduce someone representing the users.
And what ZeroUm said sounds really nice too. I thought of this earlier, but I don't know if this is possible.
Hey Drob, if you want I can come over and become the king of your blog. You've got a week to decide. I wanted to post this notice on my own blog, but nobody reads it.
I like gentoo but fear monarchies. Theres too many metaphors being thrown about.
Like, lets get fixin on this gentoo problem!
Again Daniel Robbins thank you for every thing you are trying to do. I will be installing Gentoo on my personal server
first and then my laptops. Reading your blogs and seeing your willingness to help gives me hope things can return to the glory days. One more thing. If it doesn't work out with the Gentoo Dev's or Foundation. Please folk this project I am positive you would have many followers. You could count me as one.
@wiktor: thats what the sunrise overlay is for portage.
I think Daniel summed it up nicely. There is definitely a disconnect between developers and users. I find that quite unfortunate because there is a large pool of capable users that care about particular ebuilds and would help assist the developer w/ bugs and QA.
There needs to be a way for users to easily assist developers in the ebuilds and packages they care about. One suggestion would be to allow a developer working on say packages x,y &z to cultivate their own team of users/potential developers for writing, testing & QA of ebuilds x,y&z. Instead of random joe user filing a bug on bugzilla.
That would make it easy for users like myself (who don't have time to go through the hoops to become a developer but have expertise w/ certain packages) to assist in the stability/quick release of packages I rely on. That will create a steady flow of communication between the users and devs and create a better quality ebuild faster.
Gentoo is a great tool and I hope this "crisis" marks a turning point that will make Gentoo commercially viable. Those are my two cents anyway.
Well, I must say, that things start to get interesting. Daniel, you're giving us hope for a better (Gentoo) future. :-)
I wasn't sure you would make this step, Daniel, but I'm glad you did it. Every user feels that he/she has something to say these days and this reminds me of a messy and badly done revolution, but was there a better way of doing it? I don't know.
daniel,
please fork it!
If they really give it back to you, they will not see it that way. They will "let you come back" and you will have a hard time changing things.
two important points which a lot forget is:
users of gentoo-based projects (such as SabayonLinux) are often (always?) treated as second-class "citizens"/users
and that both (esp.) by the devs and the users
the state of Gentoo (as a distribution / a whole) needs to stabilize for companies being able to work with it (again), e.g. zonbu
modularization of gentoo core-parts are already taking place: see baselayout aka openrc
Second that.
Please Daniel, fork it! Fork it and don't make the same mistakes again.
Gentoo used to be great. It is still good, and will stay good. But it will never become great again. Politics has turned it into a sinking ship.
Seems to me like a new project would bring with it more than 10% of the devs and more than 90% of the users. Not a bad start after all
I think your description of Gentoo as a cathedral is apt. Gentoo has become an oligarchy that only produces when they really need to - cite: packages.gentoo.org after two months finally being put up again and it looks to never gain a search function, the gentoo.org website finally showing an update but only after a serious fire has been set under their feet, on and on... I wrote to gentoo-nfp@gentoo.org and received a length replay to my new leadership request which I appreciated quite a bit but goes to show the state of Gentoo these days. Alot of talk, planning, hypothesizing and criticizing. I think if you have not received a consolidation by now you probably will not get one. I think another short deadline is fair but final. With the papers being delinquent, I'd look into possibly taking over the organization again. I suppose forking is possible too but initial development cost would require a considerable investment. I hope the administrators of Gentoo have the ability to come out of the cathedral and open the doors up to the public and do what's best for the community.
I think your description of Gentoo as a cathedral is apt. Gentoo has become an oligarchy that only produces when they really need to - cite: packages.gentoo.org after two months finally being put up again and it looks to never gain a search function, the gentoo.org website finally showing an update but only after a serious fire has been set under their feet, on and on... I wrote to gentoo-nfp@gentoo.org and received a length replay to my new leadership request which I appreciated quite a bit but goes to show the state of Gentoo these days. Alot of talk, planning, hypothesizing and criticizing. I think if you have not received a consolidation by now you probably will not get one. I think another short deadline is fair but final. With the papers being delinquent, I'd look into possibly taking over the organization again. I suppose forking is possible too but initial development cost would require a considerable investment. I hope the administrators of Gentoo have the ability to come out of the cathedral and open the doors up to the public and do what's best for the community.
Daniel,
Can you give us a daily update please ?
Cyrius
I start to use gentoo daily at the beginning of 2007, and i am agree with you.
I have some knowledge, but i never had the idea to help gentoo because of the cathedral :
When I come to the website, i see 'devs required' but nowhere to become one of them. User are sollicited to post bugs or improvement into bugzilla ok , but just to help the 'dev team'. In result, the 'dev' term in gentoo project mean 'code guru' for me : I don't see a "We Need YOU" panel but a "Help the Devs".
So i still User, see Changelogs, news ebuild when i Sync but i dont see a gentoo user community.
When i search docs, how-to or other I found there into wiki. Gentoo have also docs, but if i want to help (create french docs for example) i have no hope to help the "Documentation Team" : this term and organization of gentoo contributors is so far of me !. Theses wiki is an example of Users which want to help gentoo but see the hole between the two cultures.
I'm not good to arguate in english because my vocabulary is more technical oriented, but is the first time i think i could talk with the same weight of all.
So thanks to think to users...
A little (french) user like many other
Daniel,
It seems unlikely that your offer will be accepted. This saddens me greatly. Not because I wish to place you upon a pedestal as the great "Saviour" of Gentoo, but because I don't think the current devs of Gentoo really understand what is going on here.
I have carefully read what the devs have written in response to the uproar you triggered with your offer to resolve Gentoo's legal problems. And from what i am reading the devs view these issues as simply minor non-technical details. From their view point once these non-technical details are resolved everything will simply to return to status ante.
Well I for one am no longer willing to ride along with the status quo. Gentoo has gotten decidedly worse in the last few years on a number of fronts, notwithstanding the progress made on other fronts. Yet the progress on some fronts has not offset the major setbacks on other fronts.
I started with Gentoo back in 1.2 days and boy was that a lot of fun. Gentoo hasn't produced a release which I could use for my machines in the last 2 years-forcing me to use sabayon to jump start a system. And the sirens song of south african winds grows louder each day.
But far more important that the technical failings of Gentoo in the past 2-3 years is the fact that user-dev relations have utterly collapsed- the disjunct between the two is now so large that there does not even seem to be a Gentoo community any longer. Back in the day when the community feeling was strong I wanted to contribute in the small ways which I could- whether in the form of bugs, documentation or in the forums.
I stopped contributing when cianarm started stirring up troubles in the forums, and since then I have found it very, very difficult to bring myself to contribute. Perhaps this incident was just my loss of innocence, but it was also the point when I started losing faith in the community-with people like him having a say in the dev community us users have been relegated to outsiders beholden unto the whims of devs who do not care about us.
The sense of community, the sense that we are all working together for the greater good of all, is that which lowers the threshold barriers between devs and users-devs get more help and feedback from users and users contribute more to the community. In the absence of this the devs are left alone, and they feel alone, as in "us again them". This same absence hinders users from contributing time and energy.
It is not the fault of the devs that GWN died, it is a result of the lack of contributions from users in the form of material which led to its stagnation and death. It is not the fault of devs alone that know releases were made, but rather that releng did not have enough trust in us users to help them work out the bugs they were encountering, they could have elicited the creativity of the community at large to resolve the issues-but they did not.
Gentoo went from being bleeding edge to being Debian stable +6 months. And this balkanization of the community is something I have referred to repeatedly as the debianization of Gentoo-a truly saddening development. The forums are not even remotely the friendly helpful place it once was -now factions of users who identify themselves with dev cliques polarize discussions and spread derision.
I truly appreciate your courage to try to offer to help in this situation. Your name has been dragged through the mud, with continuous insinuations about your intentions and regurgitated word-of-mouth attacks surrounding your original departure from Gentoo. Daniel when we learned of your financial plight prior to your departure from Gentoo we flocked to the forums in the thousands to donate monetary and moral support. That was community. I am sure that you can see the mistakes you made back then -but the community was not interested in pointing fingers and attributing blame-we were saddened to see you have to leave.
After you left Gentoo still had enough momentum to keep going really good for another 18 months, but eventually the headlessness of Gentoo became evident. Dev in-fighting, total collapse of user dev relations, critical technical failings, and Gentoo totally losing its bleeding edge property. Most of the people I personally know who used to use Gentoo left 2-3 years ago, on to greener south african pastures.
I am not in a rush to make a decision but it has become clear to me that either work will be to bridge the gulf between users and devs and recreate the community feeling again or I will be forced to search elsewhere for a community to which I wish to belong.
I have not been able to bring myself to endorse Sabayon-even though I use it. There does not seem to be enough "critical mass" behind it to keep it moving. If Gentoo rejects your offer I hope that you would consider joining Sabayons effort and maybe enough users and devs would follow to allow Sabayon to become what Gentoo once was. Note I am not encouraging you to do some kind of fork- I understand that you would only have a certain amount of time to dedicate to such, and I certainly would never want you to find yourself in the same kind of predicament again like that which led to your original departure from Gentoo. Yet you seem to have a grasp of the issues at hand and I believe you can have a unifying effect, something which bodes well for a recreation of the community.
Today is the last day of the offer.
So? Any news? Do you discuss the whole thing with trustees or just waiting their decision?
Something needs to be addressed about how major system changes are done. I just read a post that ticked me off. I'm a user, I have a business to run. I do not hang out in the forums or read any blog with frequency. When the Gentoo devs decided to kill off kernel 2.4 I didn't know. Half of my gnome builds got unmerged before the problem was discovered. That is a production system and it is still not completely recovered. The same has happened twice in the past with Apache.
Anytime some major change that corrupts the flow of processing takes place the devs point to 'upstream'. Sorry folks that doesn't cut it. People use Gentoo because of seamless versioning. Not because it's 'easier' for you to maintain. Without users your volunteerism would be pointless. Your satisfaction of giving would be ended.
I've been very concerned for more than a year that Gentoo is coming unglued. Daniel, I hope you can help. Conceptually, Gentoo is a great distribution, but there are some things that just have to work. Some users don't always have the time searching for hours to fix things so Emerge --update needs to work everytime and snapshots, tarballs etc. must be current. And then don't leave us with crap to work with; if the LiveCD is buggy, get rid of it and focus on the minimal install. Also, the idea of new users installing Gentoo from another distribution is treasonous.
Anyway... I'm sure you know all that.... so, this user will welcome you back. And by all means, find someone to keep the front door current at www.gentoo.org.
thanks
"A user mentioned git as a repository ... that idea is a no no. Git is really slow. It is one of the worst versioning systems out there."
Have you even used git?
To quote you: "The Foundation has essentially no voice, the Council only communicates with fellow developers"
This is due entirely to the trustees lack of action. Had they any issue with anything the council had done, they didn't bring it up. The council is in charge of technical matters(and I'll refer you to last years logs when flameeyes tried to do something that was a foundation role, and I stopped him)
Last years council would have gladly worked more with the trustees had they seemed even remotely interested. We even had a trustee as a council member!
I think you're extremely biased towards some concept of a foundation that doesn't exist and are putting blame where it doesn't belong.
I have been using Gentoo for four years now, before that (and still now) I used FreeBSD. The reason I started using Gentoo was that it reminded me of FreeBSD in some ways, it was fast, stable and had a great package manager system. I loved the way everything was centralized, the way the apps were compiled to run faster on my system. There were so many apps in portage, I always found everything I needed! I loved the way I didn't have to add dozens on extra repositories like you have to in debian and red hat based distros...well, I am a user, yes, I am a developer also (not a Gentoo developer) and I love Gentoo. I have noticed that things have started getting strange, I started having to create ebuilds very often and I wondered why? I found the reason, it's the overlays, I was horrified, now gentoo's just like debian based distros, or worst, red hat based distros, where you can't find your apps in a central place so you have to add all there extra overlays and then have the same package twice, often incompatible versions, etc. You have to search in what overlay is this package that I'm looking for... I hate the overlays, I preferred the bugzilla method : if an ebuild does not exist, a user adds it to bugzilla and a dev will put it in the MAIN portage tree. I also understand that maybe this was time consuming for the devs, but maybe another alternative could have been found, like having special users that have greater privs that normal users but that don't have to go through the complicated "becoming a dev" thing, and those users goal could have been to check the ebuilds and approve them and either pass them on to the devs to be included (they wouldn't have to do any checking) or commit them themselves. I hate this whole overlay fad!
Gabriel
I have been using Gentoo for four years now, before that (and still now) I used FreeBSD. The reason I started using Gentoo was that it reminded me of FreeBSD in some ways, it was fast, stable and had a great package manager system. I loved the way everything was centralized, the way the apps were compiled to run faster on my system. There were so many apps in portage, I always found everything I needed! I loved the way I didn't have to add dozens on extra repositories like you have to in debian and red hat based distros...well, I am a user, yes, I am a developer also (not a Gentoo developer) and I love Gentoo. I have noticed that things have started getting strange, I started having to create ebuilds very often and I wondered why? I found the reason, it's the overlays, I was horrified, now gentoo's just like debian based distros, or worst, red hat based distros, where you can't find your apps in a central place so you have to add all there extra overlays and then have the same package twice, often incompatible versions, etc. You have to search in what overlay is this package that I'm looking for... I hate the overlays, I preferred the bugzilla method : if an ebuild does not exist, a user adds it to bugzilla and a dev will put it in the MAIN portage tree. I also understand that maybe this was time consuming for the devs, but maybe another alternative could have been found, like having special users that have greater privs that normal users but that don't have to go through the complicated "becoming a dev" thing, and those users goal could have been to check the ebuilds and approve them and either pass them on to the devs to be included (they wouldn't have to do any checking) or commit them themselves. I hate this whole overlay fad!
Gabriel
Some of you people are still forgetting why DR quit and went to M$. He was in debt and working on Gentoo full time. The man needed a paying gig. Getting sick of people acting like he just abandoned Gentoo for no reason. He did try to set it up to stand on it's own legs. It isn't his fault the trustees didn't follow through.
If DR learned any lesson after leaving to fix his financial situation it was not to do it Full Time anymore. So let the man come back and fix what the Dev's don't want to do.
Oh, mentioning about Dev's not letting users know what they are working on,.... One really stands out in my mind that does communicate and that is Flameeyes. Too bad he is against Daniel's return. The funny thing is that he don't like to get involved in the politics and just wants to work on stuff and make it right. So why not DR take care of the stuff F. doesn't want to do. ?
It's that "tough crap, if you don't fix it yourself then don't complain, go back to Windows" response to yet another recent instance of system update breakage that has caused me to just install Ubuntu and leave Gentoo after more than five years. It is sad that it came to this point, but most developers have long since stopped even pretending to give a crap about the users, even about such fundamental things like whether their computers even run, to say nothing of many developers' social skills in the IRC channel (the term "inbred" comes to mind).
Gentoo is dead. Long live Debian.
I wish to point Mr. Robbins and anyone reading these comments to a guest post by a Matthew Summers on Joshua Jackson's (tsunam) blog also syndicated on planet.gentoo.org.
the link:
http://tsunam.org/2008/01/21/guest-post-matthew-summers/
Please comment.
Daniel:
I've been a longtime user and spread the use of Gentoo to places like the US Army, where it has performed in excellent fashion. At least until the breakages of the last couple of years. Many systems have been removed from service and/or replaced by Windows or RHEL as a result.
Even at a personal level i'm pretty much ready to abandon Gentoo for my personal servers. These are only a few, but i'm sure i'm not the only one feeling like this: i know many people who have given up long before me, in fact.
I trace the problem to your departure from leadership. Things just have not been the same since then. I believe if your offer is not accepted then I have to move on as well.
Thank you for all of your efforts and the great product you produced.
-KP
In case someone reads till down here (I stopped about halfways - no time).
I have been a user for a few years now, and I generally liek Gentoo very much.
I am not very active in the forums (not time for another forum) and I don't read gentoo mailing lists, but when I see a problem, I try to help to solve it.
What I feel as problem is, that I often don't know how to contribute.
To state this with an example:
I am getting more and more into Python, at the moment, and I see that python 2.5.1 is still being marked as testing.
In the bug, there's been a long discussion about some applications which get broken by it, and several people proposed that they'd patch them, if they get told what exactly those packages are.
http://bugs.gentoo.org/show_bug.cgi?id=148333
There are two open dependency bugs left, and both of them have a fix in their description, so I don't know what I could do to help mark python-2.5 as stable.
How can I join?
More exactly: Where can I search for information how to join a specific gentoo subgroup?
Where this problem seems even more apparent: I see many useful python programs missing from the tree. Now there's an ebuild creating program (g-pypi), but it doesn't work on my system.
How do I help best? Where do i find the python-herd?
And how could someone who doesn't ask in here find the python-herd, or any other herd?
I don't want to have to ask. I want to be able to search and find the group, without having to ask where and how to search for it, first.
In case someone reads till down here (I stopped about halfways - no time).
I have been a user for a few years now, and I generally liek Gentoo very much.
I am not very active in the forums (not time for another forum) and I don't read gentoo mailing lists, but when I see a problem, I try to help to solve it.
What I feel as problem is, that I often don't know how to contribute.
To state this with an example:
I am getting more and more into Python, at the moment, and I see that python 2.5.1 is still being marked as testing.
In the bug, there's been a long discussion about some applications which get broken by it, and several people proposed that they'd patch them, if they get told what exactly those packages are.
http://bugs.gentoo.org/show_bug.cgi?id=148333
There are two open dependency bugs left, and both of them have a fix in their description, so I don't know what I could do to help mark python-2.5 as stable.
How can I join?
More exactly: Where can I search for information how to join a specific gentoo subgroup?
Where this problem seems even more apparent: I see many useful python programs missing from the tree. Now there's an ebuild creating program (g-pypi), but it doesn't work on my system.
How do I help best? Where do i find the python-herd?
And how could someone who doesn't ask in here find the python-herd, or any other herd?
I don't want to have to ask. I want to be able to search and find the group, without having to ask where and how to search for it, first.
To be fair you should also criticize the "take all, give nothing" mentality of a lot of users. This is not something unique for gentoo but for many popular open projects. And after reading some bug reports/ bug comments on gentoo I can understand the particular "disconnection" between devs and users.
Actually every user gives back what he thinks is apropriate. It is our fault that we expect back documentation, bug reports and code patches.
I feel like other, alternative contributions are largely underpromoted:
graphic art, music, sounds. These are now mostly done by programmers who would better use their time to what they do best - program. That is not to say programmers are not good graphic artists, but that there should be different, more focused roles. Nobody denies programmers fulfilling both roles, but there should be the possibility to do them separately.
Good post.
Post a Comment