Problems at Gentoo

From Gentoo Linux Wiki

[edit] Introduction

[edit] Why I(Slalomsk8er) Started This Wiki Page

Let me put it this way: the current event of revoking the charter of the Gentoo Foundation by the state of New Mexico, showed me that I need to do more then just whine and AllenJB's post showed me that we are "stupid" and uninformed as single parts of the community (yes he has a valid point with this) but I hope we are not a dumb mob as a whole but a little bit more then our parts combined.

So I started this page at this community tool to list things that go wrong in the eyes of the community and what we think could be done by our leaders to fix this. It shall not be a way of pushing someones agenda but to show what is going on in the negative for those that want to open their eyes and look.

[edit] What This Page is All About

The Ideas and "problems" that are listed here are not to be considered authoritative, just representative of views of the small portion of the userbase that reads and contributes to this site and how some persons think that it would be able to fix them.

[edit] Resources and Recommended Reads

• The discussion thread at this page is build on: Problems at Gentoo Discussion
• My initial announcement of this page at the "Should the Gentoo trustees accept Daniel Robbins offer?" thread: let us do some thing usefull
• Gentoo Projects and subproject status from the developer mailing list
• archives.gentoo.org mailing list archives

[edit] The Problems

[edit] Too many users reading ill-founded nonsense and providing inane opinions

[edit] Why this problem exists

• Certain users have this insane notion that they, rather than the code, are what's important about the distribution.
• Certain devs have this insane notion that they, rather than the distribution, are what's important.
• Certain users like to think that they're being helpful by discussing things, but few if any of them have the slightest clue what they're talking about.
• Certain devs wrote software which for a user is useless but this devs insist on they view. Certain users think that this is a bad thing and rather than going to use a distribution created for someone of their skill level, complain instead that the wrong tool for the job they're doing doesn't do the job they're doing properly.
• Both groups of users (devs / enduser) allowed that a deep rift could be arise. Users continue to use a distribution that's not suitable for them, whilst complaining that developers are developing things for the distribution's target audience rather than them.
• The perverse notion that 'elitism' is a bad thing. The alternative to elitism is mediocrity.
• People don't understand the Council / Foundation separation.
• Some people are under the impression that Gentoo's early success was because of drobbins, not despite drobbins.
• Most users don't know about drobbins' repeated attempts to take Gentoo commercial, use Gentoo money to finance himself and use the Gentoo branding for personal profit.
• Most users don't remember that drobbins' previous stints with Gentoo ended less than amicably due to fundamental philosophical differences with Gentoo as a whole.
• The idea that forums and a wiki are a good way of delivering documentation to end users. Often this is because of confused ideas along the lines of "wikis are like open source" (untrue, since unlike open source there's no authority doing quality control) and "wikis are democratic" (debatable, and in any case democracy requires a well informed electorate, which most wiki and forums contributors aren't). In reality, wikis and forums end up delivering *extremely* bad advice that will frequently end up crippling systems, since there's no-one knowledgeable around to stop people posting something that "works for them" as a howto (examples: TIP Safe LDFLAGS, Safe Cflags, TIP Exclude categories from emerge sync, TIP Link Root To Main User).

[edit] Possible fixes for the problem:

• This Wiki page as it was designed by me (Slalomsk8er) to give the users a tool to help them selfs in finding the knowledge it takes to do better.

[edit] Devs, Timelines, Communication or Gentoo management

thanks to bmichaelsen This I (Slalomsk8er) believe is a meta problem child problems are:

• Developers are retiring faster than they respawn
• and many more

[edit] why the problem exists:

Gentoo's leadership is lead by devs selected by devs. Those leader tend to have great technical merit but often lack in organisational skills (timelines, priorities). This sometimes results in taking no (visible) action until a near perfect solution is produced. This is a good for the long therm design of systems, but sometimes a using a "good enough" solution is better than none or a near perfect solution.

There's no alternative to the current Gentoo -- seesaw

[edit] proposed solution:

• seperate the technical leadership from the organisational. Trustees, PR, DevRel, UserRel face tasks that are no good match for the skillset of an average gentoo dev geek (no offensive, I am not too well suited either). (see http://forums.gentoo.org/viewtopic-p-4718339.html#4718339 for the origin of the idea).
• use a pro-bono administration. In Germany there is http://bodensee-consulting.org/ for example. Equivalents in the US?
• create a new codex for devs / people in general who want to work @gentoo and enforce it (we need a leader / council for that: in one word: drobbins (?)) -- kernelOfTruth
• Change the way the Gentoo management works
• Fork it ASAP -- seesaw
• use http://gen3.org
• use http://www.sabayonlinux.org/ as a alternative to a fork or see it as a slow fork

 
This new Sabayon base install is an opportunity to innovate and fix some things that you may not like about Gentoo - both the technology and community.
With sufficient user involvement, it can and will reach critical mass.
  

• Implement Team Syntegrity® like the martial arts school I train at did

 
No human network can operate faster or better

In three to four days, Team Syntegrity® will create the widest possible understanding and interconnection
within a large heterogeneous group to enable a complex, difficult and/or conflict-ridden problem to be solved.
  

• mirror the mailing lists threads in the forum to make it a one stop for the users.

[edit] Developers are retiring faster than they respawn

[edit] Why this problem exists:

 
Gentoo devs appear a secretive group that looks down on Gentoo's (often highly computer literate,
to the point of being devs themselves elsewhere)
normal users instead of being welcoming and tolerant of newbie errors.
This makes potential devs feel that they are not wanted and/or would not belong.

That's the appearance from the outside anyway and not meant to refer to any particular person.
At one point a couple of years ago I entertained for a short while a thought
that I might become a Gentoo dev but decided against it.
Now I'm committed to a completely different project.
  

 
Explanation: During the last three weeks only two new devs have joined.
During the same time 19 devs have retired.
That's an outflow of 17 devs, and if it continues at this rate, it's pretty alarming.
  

(according to this bugzilla search for the time period, I counted only the FIXED bugs)

 
Ehm, have a look at the bugs you posted: most of them are being retired due to inactivity.
Every once in a while a bunch of inactive devs is being retired, nothing unusual about that.
I still remember the 'mass-exodus' of devs a while back; this is nothing of the sort.

[edit] Proposed solution:

• Greater care taken of potential devs (currently users).

[edit] Too many ebuilds marked unstable, or no ebuilds at all

Thanks to Beetle B. his original post can be found here

[edit] Description of the Problem:

Somehow, Gentoo got behind on providing ebuilds. I've seen cases of fairly popular packages having their latest versions marked ~x86 even though a year has passed since their releases.

Worse, a number of fairly popular packages (or their latest versions) aren't even in Portage - not even masked or keyworded. Again, this is often months (if not over a year) after they've been released.

I really wish I had concrete examples to give you, but haven't been noting them down.

When I installed Gentoo back in 2003, it was known for providing ebuilds very soon after a package's latest release. If you couldn't wait, you just put it in package.keywords (OK - that system wasn't available then - but regardless, you could install it). For me, it's scary how often over the past year I've had to go to an overlay to get a much desired package.

This could be a related problem to 2.3 Developers are retiring faster than they respawn in the sense of: we need more developers for the maintenance of ebuilds.

[edit] Why this problem exists:

I don't know. I'm hoping someone here can tell me. I've thought of a number of possibilities:

1. Fewer (or the same number of) devs available to provide stable ebuilds. 2. Linux has really grown since 2003, and it's hard to keep up with all the packages out there. 3. Something else (you tell me). 4. Developers are slowly starting to actually test things before marking them stable. The move away from maintainers stabling to arch teams stabling, and the introduction of an x86 arch team, means more testing gets done so the frequency of broken things being marked stable has decreased somewhat. 5. Users simultaneously expect KDE 4 to be stable right now and to have stable actually work. 6. The increased complexity and modularity of upstream packages does not work well with the ebuild format or with Portage.

Option 2 scares me if true. Most of the complaints have been on issues involving organization, the Foundation, etc. This would be a tough technical issue to handle. Will Gentoo at one point just provide ebuilds for the basic packages and leave everything else to overlays?

It would be really nice if the devs can explain the lack of ebuilds. I don't mean this to be a complaint, and nor am I considering leaving Gentoo. It may have lagged behind on providing packages, but it still excels in other things - it's a darn good system.

 
Beetle B,

The concept and reputation of overlays has changed over recent years. They have become more respectable. (Thats not the right word but it will do)

Many herds to their development work in overlays with the result that packages get added to the tree and go to ~arch without being hard masked for development.
I see this as mixed blessing. On the positive side, it makes it easier for users to participate in testing. No more put the ebuild into your own overlay and keyword it. Just fetch the overlay with layman and use it like a part of the portage tree.
On the downside, testing packages take longer to reach the 'official' tree.

Its just a different, possibly more flexible way of working.

I'm sure you can find some examples of what you cite by looking at bugs assigned to maintainer-wanted in bugzilla be the reason is self explainatory there.
  

[edit] Possible fixes for the problem:

• Hard to say without knowing the actual problem...
• redefine package quality guidelines and the workflow to archive the stable status
• Project Sunrise - Gentoo User Overlay

 
Maybe you know this project, we call it "Sunrise".
There you can make your own steps to get a feeling of what it requires to properly maintain an app with gentoo (including getting bug reports,
suggestions from other people,...) without the need of being through the recruitment process.
Currently we have around 170 committers, ~90 who came back more than once and made 8 pretty active devs out of it.
There are around 400 ebuilds in there, some come in, some move over to portage tree, so it's a pretty constant flow.
As you see, a bunch of people seemed to have thought about it and now got involved to help out. 
  

[edit] The Bad Shape of Portage

http://planet.sabayonlinux.org/?p=105#comments <RANT ON>

       * Most of the Gentoo ebuilds don’t have the two lists split (when it is needed) causing A LOT of issues for binary packagers (which also makes Portage/Gentoo to look bad)
       * Sometimes, when they have it, they are wrong
       * There are less than a lot but more than a few ebuilds whose dependecies are really, _really_ incomplete or horribly listed/fscked up
       * There are TOO MANY USE flags while instead packages SHOULD BE split
       * Localizations should be split too, having to recompile the whole openoffice just to add one more language is quite RIDICULOUS
       * Even if I found a lot of these bugs daily (2-3 new, every day), there’s nothing I can do to fix them without wasting a lot of time on bugs.gentoo.org trying to convince developer X to fix it.

</RANT OFF>

I am not joking nor I am happy to say that. But to me, the situation is quite dramatic. Sorry, but after 7 years of Gentoo, I am quite pissed off. What I am asking Gentoo Foundation is, let me fix them. There’s too much bureaucracy, really. Will they ack. me? I don’t know.

[edit] not yet written out problems

• No Universal-CDs, but useless and unloved LiveCD with a buggy Installer and no Stage3 for Networkless-Install
• No roadmap, no project goals or other improvements
• Online Portage not active, currently nearly unusable
• No leadership through the gentoo board/council viewable for me???!!!
• The website redesign project is...?!
• Take any reason you want, the only good one in the last to years is the new baselayout-2.0 (which is still not implement as stable)
• Current foundation structure (board/council/electionperiod) seems not very effective
• We are still called ricers
• revdep-rebuild is broken http://bugs.gentoo.org/show_bug.cgi?id=41495

[edit] fixed problems

• No news since October, no real news about the project or the lose of the foundation-status

[edit] Gentoo Foundation's Charter Revoked

[edit] Why this problem exists:

The Gentoo Foundation no longer exists because the trustees didn't do their job of filing some paperwork the state of New Mexico requires to be done from every non profit foundation (on a regular basis?)

Mostly this is because of New Mexico's residency requirements, and the lack of people in New Mexico.

 
It recently lapsed, however, since required paperwork and fees were not filed on time.
  

NMSCC 2463313 has 2 public instruments

It appears that the paperwork was for 501(c)(6) status. -- Nerdanel

 
SCC Number: 	2463313
Tax & Revenue Number: 	
Incorporation Date: 	MAY 28, 2004, in NEW MEXICO
Corporation Type: 	DOMESTIC NONPROFIT
Corporation Status: 	LICENSE WAS REVOKED...NAME UNAVAILABLE
  

 
Gentoo is not a 501(c)(3) organization but is applying for 501(c)(6) status instead.
Therefore, donations are not and will not be tax-deductible in the U.S. In other countries,
the laws will differ -- please check with your lawyer.
  

Wikipedia about 501(c)(6)

It seams as if the trustees were not able to decide what to do out of 3 option.

 
OK.  As I see it, we don't even have enough people *running* to be a
trustee.  So, what should we do, instead?

Well, how about this?  First, hold an election with a few items:
- Join the SFC
- Become a 501(c)(3) in DE
- Become a 501(c)(6) in DE
  

And finally, the last one who cared but never got an answer

 
> We filed back in July/August.  No clue what's come of it, since.

Where in New Mexico? Most public records like that are online. Believe
it or not most states at least in my experience are very efficient on
processing that paper work. I would imagine it would be on file by now
no? If not almost immediately after being filed.

Not trying to gripe, bitch, blame or etc. I think follow through is a
better way to express my interest. Just curious if it's done, not that
it's my place or etc.
  

[edit] Trustees

 
4.  The Board of Trustees

Current Interim Board Members

(alphabetical order by last name)

    * Michael Cummings, mcummings
    * Chris Gianelloni, wolf31o2
    * Grant Goodyear, g2boojum
    * Renat Lumpau, rl03
    * Paul de Vrieze, pauldv
  

[edit] Michael Cummings

 
Trustee duties remain, at least until the next election,
but I let the perl team folks know earlier this week that I was going to be going offline for a little while.
Bah. Just sharing in case anyone has any burning /msg’s waiting to fire at me the second I hop on irc :)
  

 
So I once again, and this time with finality, bid Gentoo a farewell.
It’s been a fun ride, and I hope to still be able to talk to the folks that have become my friends over the years,
but I just don’t have the time and will to keep it going.
  

 
But when they cut me off from being able to retrieve trustee news/mail/info
(despite technically being a member of the trustees until replaced),
I didn’t put up a fuss but happily got on with my life.
Now it looks like the conservancy deal was never finished,
which would indeed leave Gentoo in the state that Daniel asserts.
  

[edit] Chris Gianelloni

 
Trustees: I retired as a Trustee since there's not much point without a
Foundation to run, leaving us with one (or possibly two) trustees. 
  

[edit] Grant Goodyear

[edit] Renat Lumpau

[edit] Paul de Vrieze

Last thing heard of him in the mailing list about the issue.

 
It is indeed a problem. I might clarify my position a bit. Basically I'd be 
happy to watch things going along and giving advice (more like an advisory 
board), but I don't really have the time to do the paperwork kinds of things. 
Those are also very hard if you're not in the US. In some way it is also a 
bit of knowing where to start.

I don't think it is urgent that I'm replaced, I'll still be around although 
I'll probably not be able to always reply the same day. It is just that I 
think others could do it better than I do.

Paul
  

[edit] Why this problem is largely irrelevant:

The Foundation has nothing to do with GWN, releases, keywording, technical management, running Gentoo or indeed anything relevant to the end user.

The Foundation has two purposes. First, to handle the money. This is irrelevant since almost everything useful is run off donated hardware rather than donations being used to pay for hardware, hosting etc. Secondly, to sue anybody who violates the Foundation-held copyrights. This is irrelevant since there's no way that's ever going to happen.

[edit] Sources:

http://blog.funtoo.org/2008/01/and-it-gets-worse.html http://blog.funtoo.org/2008/01/here-my-offer.html http://forums.gentoo.org/viewtopic-p-4721450.html ARTICLE 8 Nonprofit Corporations

Some Information about the Foundation: from http://forums.gentoo.org/viewtopic-p-4721786.html#4721786

Gentoo Foundation Charter ( http://www.gentoo.org/foundation/en/ ) Social Contract ( http://www.gentoo.org/main/en/contract.xml ) Proposed Bylaws of the Gentoo Foundation, Inc ( http://www.gentoo.org/foundation/en/bylaws.xml ) Gentoo Foundation Funding and Expenditures ( http://www.gentoo.org/foundation/en/funds.xml ) Gentoo Foundation Inc. Articles of Incorporation ( http://www.gentoo.org/foundation/en/articles.xml )

GENTOO FOUNDATION, INC. ( http://www.nmprc.state.nm.us/cgi-bin/prcdtl.cgi?2463313+GENTOO+FOUNDATION+INC ) --> " LICENSE WAS REVOKED...NAME UNAVAILABLE" http://linux.slashdot.org/article.pl?sid=08/01/12/0152208

[edit] Consequences

 
US Copyright law would say some stuff is Daniel Robbins' and some stuff
is now Public Domain.
  

You get things like this:

 
How many gentoo developers does it take to change a light bulb? :?
  

 
Answer:

Unknown. It was estimated that a five-member committee of lightbulb maintenance would be the right number.
However two of the members resigned and were replaced with two others.
After that other members resigned and were not replaced. The remaining group pondered which wattage should be chosen for the replacement lightbulb
or if they should outsource lightbulb maintenance to an even bigger group of people who would probably have some knowledge of electric engineering and shopping among them.
In the end, nobody bought a new lightbulb even though they promised the matter of darkness would be solved within a few days.
One member of committee justified this by him living on another continent where lightbulbs adhere to an incompatible standard.
Half a year later when the lightbulb was still out most of the committee had went AWOL except for one resigned member who sincerely said he was sorry
for his inability to fulfil or make any progress towards his grand plan for the lightbulb.

So five (or seven) was not it. Other numbers will be tried shortly after the committee of building maintenance comes to a decision.
  

[edit] Possible fixes for the problem:

• Daniel Robbins's offer as founder of Gentoo to take things in his own hands with some conditions attached http://blog.funtoo.org/2008/01/here-my-offer.html
• join the Software Freedom Conservancy Invitation to join the Software Freedom Conservancy
• make Gentoo the RedCross of GNU/Linux Gentoo NPO Business Model thoughts
• Slalomsk8ers proposal: A proposal to get out of this mess

[edit] Current Status

[1] on Fri, 18 Jan 2008 14:15:17 -0600

 
This is a brief status update.  As the trustees know, we are currently
working with the attorney in New Mexico (Wayne G. Chew, Esq) who filed
the original paperwork for the Foundation.  He does not see any problems
getting us reinstated (always subject to change once he gets into it).
  

 
Instrument 3 of 3:  	2463313 - Reinstatement
Filing Date: 	05/12/2008
Record Added: 	05/13/08 at 08:21:14
Microfilm Roll-Frame: 	0000-0000
Last action: 	Added by AMP on 05/13/08
Comments:	REINSTATED BASED ON FILING OF ALL REPORTS AND PAYMENT OF ALL FEES