Some valid points, and a high ratio of ranting.

"""
please be aware that I'm planning on orphaning ZooKeeper in the next days. The
reasoning behind that will shortly be available as a video or slides[1] of my
talk at BerlinBuzzWords. - In short: The code quality is of such a kind that I
don't consider it good enough for Debian.
"""

http://lists.debian.org/debian-devel/2011/06/msg00316.html

http://berlinbuzzwords.de/sites/berlinbuzzwords.de/files/thomas_koch_zookeeper.pdf

--
Gustavo Niemeyer
http://niemeyer.net
http://niemeyer.net/blog
http://niemeyer.net/twitter

Hi Gustavo. Definitely some valid points in there, obv not everyone
agrees with all of them. ;-)

Thomas also sent some details earlier on the dev list, which you can find here:
http://markmail.org/message/4dzk5osxssaibirt

Discussing these points is on the agenda during our meetup around the
Hadoop summit.  There's already been significant progress on some of
it - for example a Maven patch is in the works.

Regards,

Patrick

On Fri, Jun 10, 2011 at 11:41 AM, Gustavo Niemeyer <[EMAIL PROTECTED]> wrote:
> Some valid points, and a high ratio of ranting.
>
> """
> please be aware that I'm planning on orphaning ZooKeeper in the next days. The
> reasoning behind that will shortly be available as a video or slides[1] of my
> talk at BerlinBuzzWords. - In short: The code quality is of such a kind that I
> don't consider it good enough for Debian.
> """
>
> http://lists.debian.org/debian-devel/2011/06/msg00316.html
>
> http://berlinbuzzwords.de/sites/berlinbuzzwords.de/files/thomas_koch_zookeeper.pdf
>
> --
> Gustavo Niemeyer
> http://niemeyer.net
> http://niemeyer.net/blog
> http://niemeyer.net/twitter
>

Hi Patrick,

> Hi Gustavo. Definitely some valid points in there, obv not everyone
> agrees with all of them. ;-)
>
> Thomas also sent some details earlier on the dev list, which you can find here:
> http://markmail.org/message/4dzk5osxssaibirt

Ah, sorry for the double post.  I got to know about it through today's message.

> Discussing these points is on the agenda during our meetup around the
> Hadoop summit.  There's already been significant progress on some of
> it - for example a Maven patch is in the works.

That's good to hear.  Given that the rant is being thrown in the wind
as it is, it'd be good to have a public reference to point people to
that addresses some of these points.

--
Gustavo Niemeyer
http://niemeyer.net
http://niemeyer.net/blog
http://niemeyer.net/twitter

On 10 June 2011 12:58, Gustavo Niemeyer <[EMAIL PROTECTED]> wrote:

> Hi Patrick,
>
> > Hi Gustavo. Definitely some valid points in there, obv not everyone
> > agrees with all of them. ;-)
> >
> > Thomas also sent some details earlier on the dev list, which you can find
> here:
> > http://markmail.org/message/4dzk5osxssaibirt
>
> Ah, sorry for the double post.  I got to know about it through today's
> message.
>
> > Discussing these points is on the agenda during our meetup around the
> > Hadoop summit.  There's already been significant progress on some of
> > it - for example a Maven patch is in the works.
>
> That's good to hear.  Given that the rant is being thrown in the wind
> as it is, it'd be good to have a public reference to point people to
> that addresses some of these points.
>

It also appears that maintenance of the Debian package will be picked up, if
I read this message correctly:

http://lists.debian.org/debian-devel/2011/06/msg00317.html

Henry

>
> --
> Gustavo Niemeyer
> http://niemeyer.net
> http://niemeyer.net/blog
> http://niemeyer.net/twitter
>

--
Henry Robinson
Software Engineer
Cloudera
415-994-6679

> It also appears that maintenance of the Debian package will be picked up, if
> I read this message correctly:
>
> http://lists.debian.org/debian-devel/2011/06/msg00317.html

Yeah, hopefully Clint can give us a hand there.  We are both focused
on Ensemble [1] which depends on ZooKeeper, and while we would/will
appreciate cleaner code and a more polished project over time, there's
no better Open Source project better than ZooKeeper in this space,
honestly.  Doozer looks promising, but it's still too young.

[1] https://ensemble.ubuntu.com

--
Gustavo Niemeyer
http://niemeyer.net
http://niemeyer.net/blog
http://niemeyer.net/twitter

On Fri, Jun 10, 2011 at 12:58 PM, Gustavo Niemeyer <[EMAIL PROTECTED]> wrote:
> Hi Patrick,
>
>> Hi Gustavo. Definitely some valid points in there, obv not everyone
>> agrees with all of them. ;-)
>>
>> Thomas also sent some details earlier on the dev list, which you can find here:
>> http://markmail.org/message/4dzk5osxssaibirt
>
> Ah, sorry for the double post.  I got to know about it through today's message.
>

No worries, although I hope you correct the record on your twitter
feed. ;-) This is one person's opinion, "not good enough for Debian"
implies more weight, more like "not good enough for Thomas". (see
Henry's response to this thread, re this already being picked up by
another maintainer.)

>> Discussing these points is on the agenda during our meetup around the
>> Hadoop summit.  There's already been significant progress on some of
>> it - for example a Maven patch is in the works.
>
> That's good to hear.  Given that the rant is being thrown in the wind
> as it is, it'd be good to have a public reference to point people to
> that addresses some of these points.

Well I hope our track record of building a solid system with very few
bugs, esp given the production level use we see (yahoo, SU, facebook,
HBase, twitter, etc...), would speak for itself but ... sure. :-)

Patrick

> No worries, although I hope you correct the record on your twitter
> feed. ;-) This is one person's opinion, "not good enough for Debian"
> implies more weight, more like "not good enough for Thomas". (see
> Henry's response to this thread, re this already being picked up by
> another maintainer.)

Yes, Clint will likely pick it up because we are using it.

My post:

    "The Debian #ZooKeeper packager is going to orphan it because the
code quality is not good enough for Debian:"

His message:

    "I'm planning on orphaning ZooKeeper in the next days. (..) In
short: The code quality is of such a kind that I don't consider it
good enough for Debian."

I don't think there's anything to be corrected?

> Well I hope our track record of building a solid system with very few
> bugs, esp given the production level use we see (yahoo, SU, facebook,
> HBase, twitter, etc...), would speak for itself but ... sure. :-)

I'm not the guy bashing, and I wouldn't stay silent if it was one of
my projects, but in this case it's up to you really.

--
Gustavo Niemeyer
http://niemeyer.net
http://niemeyer.net/blog
http://niemeyer.net/twitter

I had a long talk with Thomas at Buzzwords.

He is far more congenial in person than on the mailing lists.

I tried to talk about the wisdom of giving first and then criticizing
and the debatable virtue of stronger than necessary language.  He was
a good listener and I think he has stuff to contribute if he can see
past these (relatively minor) issues.

On Fri, Jun 10, 2011 at 10:47 PM, Gustavo Niemeyer <[EMAIL PROTECTED]> wrote:
>
>> Well I hope our track record of building a solid system with very few
>> bugs, esp given the production level use we see (yahoo, SU, facebook,
>> HBase, twitter, etc...), would speak for itself but ... sure. :-)
>
> I'm not the guy bashing, and I wouldn't stay silent if it was one of
> my projects, but in this case it's up to you really.

> I had a long talk with Thomas at Buzzwords.
>
> He is far more congenial in person than on the mailing lists.

Whether he's kind or not isn't really the concern at this point.  The
point is that there are real issues being raised publicly.  ZooKeeper
has a good track record on stability, but I have already heard the
code isn't clean or pleasant from different people.  It is also true
that some features have been taking a very long time to land. These
are pretty interesting points for a software project.  Is that the
case?  Is there an intention to clean up the code base?  Are features
taking too long?  What's scheduled for the current milestone and when
is it going to be released?

If you debate about these concerns over the phone and don't discuss
them publicly, what stays is a message on debian-devel pointing out
that "Live is too short for crap." [sic]

--
Gustavo Niemeyer
http://niemeyer.net
http://niemeyer.net/blog
http://niemeyer.net/twitter

On Fri, Jun 10, 2011 at 2:18 PM, Gustavo Niemeyer <[EMAIL PROTECTED]> wrote:
>> I had a long talk with Thomas at Buzzwords.
>>
>> He is far more congenial in person than on the mailing lists.
>
> Whether he's kind or not isn't really the concern at this point.  The
> point is that there are real issues being raised publicly.  ZooKeeper
> has a good track record on stability, but I have already heard the
> code isn't clean or pleasant from different people.  It is also true
> that some features have been taking a very long time to land. These
> are pretty interesting points for a software project.  Is that the
> case?  Is there an intention to clean up the code base?  Are features
> taking too long?  What's scheduled for the current milestone and when
> is it going to be released?

You can see our release history here:
https://issues.apache.org/jira/browse/ZOOKEEPER#selectedTab=com.atlassian.jira.plugin.system.project%3Aversions-panel

3.4.0 is definitely overdue, the move to TLP took significantly more
effort that we originally accounted for and that's slowed us down.
Mahadev is the release manager for 3.4.0 and is working to get it in
shape. (there's a thread on the dev list about this)

Regards,

Patrick

All I have to say is

  "Life is too short for crap".

We are using ZK for large-scale mission-critical clustering for
almost 9 months now, and guess what?   Zero defects AFAIC.

ZK contributors. I wholeheartedly appreciate your contributions.
Life is too short for crap.  Move on...

Chang

2011. 6. 11., 오전 9:52, Patrick Hunt 작성:

> On Fri, Jun 10, 2011 at 2:18 PM, Gustavo Niemeyer <[EMAIL PROTECTED]> wrote:
>>> I had a long talk with Thomas at Buzzwords.
>>>
>>> He is far more congenial in person than on the mailing lists.
>>
>> Whether he's kind or not isn't really the concern at this point.  The
>> point is that there are real issues being raised publicly.  ZooKeeper
>> has a good track record on stability, but I have already heard the
>> code isn't clean or pleasant from different people.  It is also true
>> that some features have been taking a very long time to land. These
>> are pretty interesting points for a software project.  Is that the
>> case?  Is there an intention to clean up the code base?  Are features
>> taking too long?  What's scheduled for the current milestone and when
>> is it going to be released?
>
> You can see our release history here:
> https://issues.apache.org/jira/browse/ZOOKEEPER#selectedTab=com.atlassian.jira.plugin.system.project%3Aversions-panel
>
> 3.4.0 is definitely overdue, the move to TLP took significantly more
> effort that we originally accounted for and that's slowed us down.
> Mahadev is the release manager for 3.4.0 and is working to get it in
> shape. (there's a thread on the dev list about this)
>
> Regards,
>
> Patrick

I am constantly surprised to hear that the ZK code isn't "clean or
pleasant". I was actually very happy when I opened up the ZK code base
for the first time at how generally readable and tested the code was.
It's a complex system, implemented over a fair period of time, so it's
inevitable that there will be some stuff that is hard to understand
but I'm not sure that the complexity is excessive or unmanageable.

We do have a problem with scheduling and sticking to schedules. I get
the feeling that maybe the project is in a state of growing pains as
the user base is expanding and many more people have pet features they
would like to add. But none of this is insurmountable and it would be
throwing the baby out with the bathwater to declare the project
unusable for those reasons.

C
On Fri, Jun 10, 2011 at 5:18 PM, Gustavo Niemeyer <[EMAIL PROTECTED]> wrote:
>> I had a long talk with Thomas at Buzzwords.
>>
>> He is far more congenial in person than on the mailing lists.
>
> Whether he's kind or not isn't really the concern at this point.  The
> point is that there are real issues being raised publicly.  ZooKeeper
> has a good track record on stability, but I have already heard the
> code isn't clean or pleasant from different people.  It is also true
> that some features have been taking a very long time to land. These
> are pretty interesting points for a software project.  Is that the
> case?  Is there an intention to clean up the code base?  Are features
> taking too long?  What's scheduled for the current milestone and when
> is it going to be released?
>
> If you debate about these concerns over the phone and don't discuss
> them publicly, what stays is a message on debian-devel pointing out
> that "Live is too short for crap." [sic]
>
> --
> Gustavo Niemeyer
> http://niemeyer.net
> http://niemeyer.net/blog
> http://niemeyer.net/twitter
>

THe complaint about multi talking a long time is somewhat misleading
as well.  I was able to complete the client and wire parts in a week
or so and then had no time for months.  Then Marshall and Camille
showed up (esp. Marshall) and things got done in very short order.

I have subscribed to the debian list and will provide a countervailing
viewpoint.  Code metrics are a method to achieve a goal that ZK has
come pretty close to achieving.  Such metrics are not a goal in and of
themselves.

On Sun, Jun 12, 2011 at 8:50 AM, Camille Fournier <[EMAIL PROTECTED]> wrote:
> I am constantly surprised to hear that the ZK code isn't "clean or
> pleasant". I was actually very happy when I opened up the ZK code base
> for the first time at how generally readable and tested the code was.
> It's a complex system, implemented over a fair period of time, so it's
> inevitable that there will be some stuff that is hard to understand
> but I'm not sure that the complexity is excessive or unmanageable.
>
> We do have a problem with scheduling and sticking to schedules. I get
> the feeling that maybe the project is in a state of growing pains as
> the user base is expanding and many more people have pet features they
> would like to add. But none of this is insurmountable and it would be
> throwing the baby out with the bathwater to declare the project
> unusable for those reasons.
>
> C
>
>
> On Fri, Jun 10, 2011 at 5:18 PM, Gustavo Niemeyer <[EMAIL PROTECTED]> wrote:
>>> I had a long talk with Thomas at Buzzwords.
>>>
>>> He is far more congenial in person than on the mailing lists.
>>
>> Whether he's kind or not isn't really the concern at this point.  The
>> point is that there are real issues being raised publicly.  ZooKeeper
>> has a good track record on stability, but I have already heard the
>> code isn't clean or pleasant from different people.  It is also true
>> that some features have been taking a very long time to land. These
>> are pretty interesting points for a software project.  Is that the
>> case?  Is there an intention to clean up the code base?  Are features
>> taking too long?  What's scheduled for the current milestone and when
>> is it going to be released?
>>
>> If you debate about these concerns over the phone and don't discuss
>> them publicly, what stays is a message on debian-devel pointing out
>> that "Live is too short for crap." [sic]
>>
>> --
>> Gustavo Niemeyer
>> http://niemeyer.net
>> http://niemeyer.net/blog
>> http://niemeyer.net/twitter
>>
>

On Sat, Jun 11, 2011 at 11:50 PM, Camille Fournier <[EMAIL PROTECTED]> wrote:
> We do have a problem with scheduling and sticking to schedules. I get
> the feeling that maybe the project is in a state of growing pains as
> the user base is expanding and many more people have pet features they
> would like to add.

Just to be clear, afaik this is the only hadoop related project that
doesn't have full time paid staff assigned to it, so we're all
volunteering personal time here. That combined with the effort of
moving to TLP status has definitely slowed us down of late. We really
do need to get 3.4.0 rolling though.

Patrick

Gustavo/others,
  3.4 is definitely overdue. I will be sending out the details and
date for 3.4 release. It should be soon (within weeks and not months).
As for code quality, I am really surprised that folks are trying to
debate about code quality of a product that has been in production for
3 years, is proven and used many big software/financial companies. I
have never seen any software been in production ofr 3-4 years have
perfect codebase. Parts of software always need re work (thats what
usually happens with software running/sustained for a while).
I am really sad this issue being brought up on dev list again. I hope
we can have more fruitful conversations than "code quality" on the
mailing list.

thanks
mahadev
On Fri, Jun 10, 2011 at 2:18 PM, Gustavo Niemeyer <[EMAIL PROTECTED]> wrote:
>> I had a long talk with Thomas at Buzzwords.
>>
>> He is far more congenial in person than on the mailing lists.
>
> Whether he's kind or not isn't really the concern at this point.  The
> point is that there are real issues being raised publicly.  ZooKeeper
> has a good track record on stability, but I have already heard the
> code isn't clean or pleasant from different people.  It is also true
> that some features have been taking a very long time to land. These
> are pretty interesting points for a software project.  Is that the
> case?  Is there an intention to clean up the code base?  Are features
> taking too long?  What's scheduled for the current milestone and when
> is it going to be released?
>
> If you debate about these concerns over the phone and don't discuss
> them publicly, what stays is a message on debian-devel pointing out
> that "Live is too short for crap." [sic]
>
> --
> Gustavo Niemeyer
> http://niemeyer.net
> http://niemeyer.net/blog
> http://niemeyer.net/twitter
>

--
thanks
mahadev
@mahadevkonar

(...)
> I am really sad this issue being brought up on dev list again. I hope
> we can have more fruitful conversations than "code quality" on the
> mailing list.

Stability and code quality are two different aspects, and I for one
think the mailing list for the project itself is the right place to
have these conversations.  You don't have to agree with the points
made, but being so offensive towards even bringing the concept up
feels a bit like validating some of Thomas' point, interestingly.

--
Gustavo Niemeyer
http://niemeyer.net
http://niemeyer.net/blog
http://niemeyer.net/twitter

I have been corresponding with the Debian developer community.  Their major
beef with ZK is that it doesn't compile on MIPS architecture using gcj.  The
package as proposed says that the default java is acceptable and on MIPS
this is gcj.

Clearly there is a major disconnect between what Zookeeper is, what it
requires and the current proposed Debian packaging of same.

My major question is: do we care?

Literally, is there anybody out there who would be positively impacted by
having a debian package for Zookeeper?  (nobody can be negatively impacted
at this point because there isn't any official debian package for ZK).

If there is somebody whom such a package would help, then I say we do care
and somebody ought to step up to maintain and fix this package.  Right now,
just restricting the package to use java 1.6 should be enough to move
forward.  Behind that issue, there is a moderate bug in the way that a cron
job is set up to clean up logs, but isn't unscheduled.  Beyond that, my
guess is that the level of effort should be no more than tracking production
ZK releases.

Is anybody out there?
On Wed, Jun 15, 2011 at 7:32 PM, Gustavo Niemeyer <[EMAIL PROTECTED]>wrote:

> (...)
> > I am really sad this issue being brought up on dev list again. I hope
> > we can have more fruitful conversations than "code quality" on the
> > mailing list.
>
> Stability and code quality are two different aspects, and I for one
> think the mailing list for the project itself is the right place to
> have these conversations.  You don't have to agree with the points
> made, but being so offensive towards even bringing the concept up
> feels a bit like validating some of Thomas' point, interestingly.
>
> --
> Gustavo Niemeyer
> http://niemeyer.net
> http://niemeyer.net/blog
> http://niemeyer.net/twitter
>

Gustavo,
 I am all for code quality. I myself am targetting 3.5.0 for some code
cleanup. The issues have been talked about and some of them will be
addressed and some other are pure nonsense. Complaning about in all
over the world doesnt help and posting those complaints time and again
on the dev list (folks who are working on it) is offensive.

I am ok with Debian not supporting ZK. I think rpm/debian packages
should be provided by the project itself.

thanks
mahadev

On Wed, Jun 15, 2011 at 10:32 AM, Gustavo Niemeyer <[EMAIL PROTECTED]> wrote:
> (...)
>> I am really sad this issue being brought up on dev list again. I hope
>> we can have more fruitful conversations than "code quality" on the
>> mailing list.
>
> Stability and code quality are two different aspects, and I for one
> think the mailing list for the project itself is the right place to
> have these conversations.  You don't have to agree with the points
> made, but being so offensive towards even bringing the concept up
> feels a bit like validating some of Thomas' point, interestingly.
>
> --
> Gustavo Niemeyer
> http://niemeyer.net
> http://niemeyer.net/blog
> http://niemeyer.net/twitter
>

--
thanks
mahadev
@mahadevkonar

>  I am all for code quality. I myself am targetting 3.5.0 for some code
> cleanup. The issues have been talked about and some of them will be
> addressed and some other are pure nonsense. Complaning about in all

That's excellent news, thanks for working on this Mahadev.

> over the world doesnt help and posting those complaints time and again
> on the dev list (folks who are working on it) is offensive.

Again, you seem too easily offended when touching on the subject,
which makes Thomas' points valid.  If you are already working on clean
ups, simply pointing out what is going on will be much more effective
and comforting than fighting about the subject being touched by people
that use and care about your software.

--
Gustavo Niemeyer
http://niemeyer.net
http://niemeyer.net/blog
http://niemeyer.net/twitter

> Again, you seem too easily offended when touching on the subject,
> which makes Thomas' points valid.  If you are already working on clean
> ups, simply pointing out what is going on will be much more effective
> and comforting than fighting about the subject being touched by people
> that use and care about your software.
>
Gustavo,
 I apologize to be so touchy on this. There is some history behind
that. I had been very willing to work with Thomas to work on clearing
out some issues but looks like instead all this happened.

As I think you would understand projects running for long time have
issues with convoluted code  in some places or the other, which as
part of sustaining need to be fixed. We are in that phase, so I am not
sure where all this is coming from.
But you are right that next time I should point to jiras. But for now,
I think we should avoid ZK packages in Debian.

thanks
mahadev

> --
> Gustavo Niemeyer
> http://niemeyer.net
> http://niemeyer.net/blog
> http://niemeyer.net/twitter
>

--
thanks
mahadev
@mahadevkonar

> As I think you would understand projects running for long time have
> issues with convoluted code  in some places or the other, which as
> part of sustaining need to be fixed. We are in that phase, so I am not
> sure where all this is coming from.

I definitely understand it, and my main concern was precisely to avoid
having an undue impression sticking.  We have packages for ZooKeeper,
have contributed patches, are using it in Ensemble [1], and will
continue to do so in the foreseeable future.

[1] https://ensemble.ubuntu.com

--
Gustavo Niemeyer
http://niemeyer.net
http://niemeyer.net/blog
http://niemeyer.net/twitter

>  Again, you seem too easily offended when touching on the subject,
> which makes Thomas' points valid.  

No it doesn't. He is allowed to be offended and it has no bearing on the truth of the matter. This isn't a Shakespearian play.

C

-----Original Message-----
From: Gustavo Niemeyer [mailto:[EMAIL PROTECTED]]
Sent: Wednesday, June 15, 2011 3:04 PM
To: [EMAIL PROTECTED]
Subject: Re: Debian packager orphaning ZooKeeper

>  I am all for code quality. I myself am targetting 3.5.0 for some code
> cleanup. The issues have been talked about and some of them will be
> addressed and some other are pure nonsense. Complaning about in all

That's excellent news, thanks for working on this Mahadev.

> over the world doesnt help and posting those complaints time and again
> on the dev list (folks who are working on it) is offensive.

Again, you seem too easily offended when touching on the subject,
which makes Thomas' points valid.  If you are already working on clean
ups, simply pointing out what is going on will be much more effective
and comforting than fighting about the subject being touched by people
that use and care about your software.

--
Gustavo Niemeyer
http://niemeyer.net
http://niemeyer.net/blog
http://niemeyer.net/twitter

Regardless of offense taken due to his approach in describing them, some of
Thomas' suggestions are reasonable (some switch statements are near
replicas, many unit tests are framed as integration tests).  Some are simply
esthetic (de gustibus and so on) and some are (IMHO) not such a great thing
at all.

Zookeeper is pretty open to positive change, but not very open to new-comers
appearing with suggestions for extensive re-writes just to please a
particular esthetic preference.  And frankly, complaining in other fora
decreases credibility with me even more than just complaining does.
 Establishing a record of viable improvements increases credibility.

And, what do you know, we just re-invented Apache meritocracy from first
principles.

On Wed, Jun 15, 2011 at 9:19 PM, Fournier, Camille F. <
[EMAIL PROTECTED]> wrote:

> >  Again, you seem too easily offended when touching on the subject,
> > which makes Thomas' points valid.
>
> No it doesn't. He is allowed to be offended and it has no bearing on the
> truth of the matter. This isn't a Shakespearian play.

hey i've put up a wiki to collect some of these issues from thomas and
others: https://cwiki.apache.org/confluence/display/ZOOKEEPER/ImplementationCriticisms

if people have others to add, please feel free to update.

ben

On Fri, Jun 10, 2011 at 11:55 AM, Patrick Hunt <[EMAIL PROTECTED]> wrote:
> Hi Gustavo. Definitely some valid points in there, obv not everyone
> agrees with all of them. ;-)
>
> Thomas also sent some details earlier on the dev list, which you can find here:
> http://markmail.org/message/4dzk5osxssaibirt
>
> Discussing these points is on the agenda during our meetup around the
> Hadoop summit.  There's already been significant progress on some of
> it - for example a Maven patch is in the works.
>
> Regards,
>
> Patrick
>
> On Fri, Jun 10, 2011 at 11:41 AM, Gustavo Niemeyer <[EMAIL PROTECTED]> wrote:
>> Some valid points, and a high ratio of ranting.
>>
>> """
>> please be aware that I'm planning on orphaning ZooKeeper in the next days. The
>> reasoning behind that will shortly be available as a video or slides[1] of my
>> talk at BerlinBuzzWords. - In short: The code quality is of such a kind that I
>> don't consider it good enough for Debian.
>> """
>>
>> http://lists.debian.org/debian-devel/2011/06/msg00316.html
>>
>> http://berlinbuzzwords.de/sites/berlinbuzzwords.de/files/thomas_koch_zookeeper.pdf
>>
>> --
>> Gustavo Niemeyer
>> http://niemeyer.net
>> http://niemeyer.net/blog
>> http://niemeyer.net/twitter
>>
>