Activity
February 22nd, 2008Gentoo has once again fallen into a rip van winkle state. For a period everyone was motivated and driven about Gentoo both in the developer and user communities. A few people tried to harness that and improve things. Donnie has probably had some of the most success with the revivial of the Public Relations group and keeping the front page updated. This has most certainly calmed one of the major issues that users had brought up about Gentoo as an Organization.
As Daniel so kindly pointed out that in a lot of ways that Gentoo is very likely a Dystopian society. Much like in 1984 you have your inner party that would be made up of your Developers, who could be split in a few ways such as by length of service, amount of work done, ebuild/non ebuild developers. Depending on how you split the above you can have Developers in the Outer party and users who make contributions as well. Finally you have the proles who are the users who use the product for their own reasons but also unlike the novel have the freedom to leave and go to another distribution that is the same as Gentoo in this respect. Ultimately, in any shared project that is Open source I tend to see the same global phenonomen that I described above with Gentoo. When you get down to individuals however is when you see a change.
Oceania(Gentoo), as we’re all aware is entirely driven by the vollunteers that join and contribute to make it better for what they need. Notice that its not typically for the betterment of the party but for their own individual needs. I’m not for example going to suddenly take an interest in clustering because I’m told that I need to. I’ll stick to what I’m doing which is a mixture of areas that I try to affectionately term as the Human Relations.
As part of the Relations aspect, I’ve been trying to get more information out to the proles and trying to request information as well. I know that the developers/inner and outer party are far outweighed by this large group and in ways I’m trying to organize at least some people who have a hint of motivation to give me ideas of things that they see. I’ve unfortunately not been successful with this task. Both here and on the forums which I’ve started posting to again after a year plus absense. I would hope that it would be agreed that I’m trying to reach out but someone has to take my hand and give me something as well. A few have done this and while I might not agree with the idea I will at least discuss it and give insight into the party because of that discussion.
I continue to ask for topics to talk about both from my fellow members of User Relations, who also seem to be busy with other things and do not simply have a few minutes to answer some questions I’ve posed to them. If I can’t even get an answer from people I should be working with, I’m not sure how effective I can be with anyone else. I’ve had a few minor successes with the request to discuss and explain what Glep-55 was about and would like to cover other things like it but I don’t know what people would like to know about.
So once again I’m here asking for people to talk to me and give me their ideas. Its not a trick to get you to commit thoughtcrime. Its simply a request from someone who wants to improve Gentoo for everyone. To have people who want to see it improve step forward with me and see if we can’t come up with ways that would help people. I still need to work on a few things naturally like a quick guide about submitting stable requests.
February 22nd, 2008 at 11:59 am
I found your blog on google and read a few of your other posts. I just added you to my Google News Reader. Keep up the good work. Look forward to reading more from you in the future.
Jason Whitmen
February 22nd, 2008 at 5:34 pm
Hi everyone, I didn’t yet use the chance to give my ideas about gentoo publicly. Tsunam’s repeated posts appearing on planet gentoo finally persuaded me to take this step. It may very well be that someone already suggested some of the thing’s I’m about to mention, so please accept my apologies if I’m wasting your time….
*** how to involve people
as far as i know, the simplest way to become officially part of gentoo is to do the ’staff quiz’. since i was/am interested, i started to work through the questions and found out that some are quite difficult to answer well. Surely, i might just guess and probably be right. But since I want my quiz answers to be 100% correct I’m painfully googling for every single one, here and there without much luck. (i hate to waste time, so i’d feel bad if my answers werent good enough on first try because if the person checking them. as i’m also selfish and want others to work on gentoo so that i can use it, so wasting the time of people working on gentoo is against my own interest). Maybe there’s some doc somewhere covering topics I found hard to answer it but since gentoo’s documentation is quite cluttered, if i cant google it i wont know. suggestion: either write some docs on gentoo or improve docs
*** how to improve gentoo’s documentation
gentoo’s docs aren’t easy to navigate. to give an example, tips on securing a gentoo box are in the handbook, in the gentoo security handbook, on hardened projects pages and god knows where else. same goes with other topics… some are in the main gentoo docs, some in specific projects’ pages. my suggestion:
a) implement a search engine (i.e. mysql/sphinx) that would allow joint docs+forums searches
b) implement a taxonomy (tagging system) that would allow gentoo staff (and maybe users too?) to tag documents and their parts with keywords. so for example, the ’security’ tag would be assigned to http://www.gentoo.org/proj/en/hardened/, http://www.gentoo.org/doc/en/security/index.xml and http://www.gentoo.org/doc/en/handbook/handbook-x86.xml?part=x&chap=y#anchor-z …. there is the http://www.gentoo.org/doc/en/list.xml , but some docs cover more then one topic and i’m not sure if that’s reflected in the list
c) implement both
*** how to get more developers
by far the biggest problem in becoming a devs is the steep learning curve for people to get accustomed with writing ebuilds. the dev handbook is so repelling and to try things out things look so time consuming that for people with less free time its just easier to give up on that. my suggestions:
a) teach by example. dissect some simple, moderately complex and very complex ebuilds to introduce the user into the problematic intuitively. a very complex ebuild might be for example php or mysql-community, i didnt dare to look at X
b) dev proxying / tutoring … allow people interested in becoming devs or at least helping out to get in touch with devs. while there’s nothing hindering us non-devs to do that now, most devs seem really busy. so if a page existed with devs and the herds they work in with i.e. green or red circles announcing the status “willing or not willing to accept a newbie to bug me with stupid questions about how to maintain a package” might be helpful. i’d also suggest a probation time that would show if the dev and the maybe-to-be-dev can go along well together since not everyone needs to be overjoyed by the other person’s personality / workstyle …
i’d have more ideas to give but it’s late here so maybe next time ..
February 22nd, 2008 at 8:03 pm
I’d guess one which which promotes activity for other distributions is the thought of a release or deadline of some sort. Take Ubuntu or Fedora for example, they’ve got fairly specific release dates and features tend to make it into the next release which everyone works towards.
Two distributions which I see with activity problems tend to be Debian and Gentoo. Debian has the problem that the next release is “when it’s ready”, and Gentoo is versionless apart from the installation download. How can you get anyone excited about the next big release when everything just moves along slowly?
You say that people join Gentoo and only improve their little area, but isn’t this how Gentoo is structured? You aren’t supposed to touch anyone elses packages and submit bug reports on them instead. Look at the last paragraph in your post, you’re going to submit a load of bug requests just to get packages stabilised.
I’d say things have become a little too complex to make developing for Gentoo fun anymore.
February 22nd, 2008 at 9:33 pm
I’m hard pressed to give you ideas to improve things. I’ve mentioned before that I wish games were supported better. I wish audio recording/sequencing work better out of the box. I’ve recently discovered that miro looks cool too. But I know that these things aren’t trivial to support, there probably isn’t big demand for them from users or developers and I know that gentoo is basically driven by a herd of cats.
So what do I ask for? I’m left feeling that the request for ideas has to also be a request for action to carry out those ideas. But alas, I am but a prole. I don’t have to time that I would have had a decade ago to be a developer or to sit in irc and comb forums and live an internet life. Frankly, it’s solely because of the rss feeds that I even have time to keep up the way I do. I have tried to do little things to contribute, but no one even has time to respond as you’ve noted. My biggest attempt to contribute to gentoo has been a much larger time sink than I intended or thought necessary and since it has been completely ignored, it seems to have been a waste of time as well. I’m not repeating that mistake. Gentoo does not seem to be open for small contributions. At least not in the areas I care about.
Ideas to improve things can come from things bug me. My first big complain I guess came when xmms died. It was because I didn’t know it was coming, I was unprepared and I am still recovering as far as my music listen habits at home go. At that time I learned I needed to watch the gentoo newsletter to be made aware when a cat got moody and decided to dump a difficult but popular package. I was fine with that. There was a last-rites list and a removed list and I learn to watch them religiously. (Another mistake I don’t want to repeat.) But then the newsletter fell off the face of the earth. Eventually it came back, but now I find (and I commented to the GMN feedback address) that packages are showing up in both last-rites and removed lists in the same issue, not really providing much reaction time. I guess now I’ll have to monitor some mailing list and/or the forums if I want this basic info. It sounds time consuming. It might be nice to make this less difficult for users so we don’t have to monitor list that we probably should be responding too.
Overall, one thing I haven’t found a way to bring up in previous comments is my general feelings about gentoo. Since I’ve started using it, I have had a few problems. The initial learning curve on the install and dealing with merge issues was a little painful. Some issues when I had to move $DISTDIR, $PORTAGE_TMPDIR & $GAMES_* due to space constraints was a little painful. Breakages when important packages disappeared or special upgrade steps (that were only mentioned in a blog or forum post) were needed was a little painful. But overall, gentoo is very flexible and it gives me the ability to configure things the way I like a lot easier than any distro I’ve used to date. Overall, when it comes to gentoo, I’ve been fat dumb and happy. There have been annoyances but they haven’t been too bad yet.
However having to monitor so much net traffic just to get basic info is getting to be more than a little painful. Add to that a number blog posts that point out that developers rule, users drool, the developers do what they want and the users can politely (or not so politely) stfu or go away. It gets me more bothered than I care to be. Whether it’s a dramatic portrayal of a Dystopian society as you put it, the harsh and to-the-point volunteer-developers-do-want-they-want approach of flame eyes or the one kind invitation to try debian from an unnamed developer, it gets rather irksome. I know it’s true and if the cats do get too moody I will probably explore the other alternatives in the world like I did when debian stagnated.
While communication is itself important, I guess my main idea might be to try to separate the basic info users need to keep the system going from some of the expressions of the developers feeling about the users. In such large doses, it’s just not good for public relations.
Edd
February 23rd, 2008 at 1:47 am
something that is really missing is positive noise.
distrowatch or pro-linux.de are providing a lot of people with news from the distro world. But the only times gentoo shows up there is when some devs started a flamewar, drobbins attacked again or something bad happened. Some dev was extremly harsh on some user? It will show up in the comments section of dw,pl.de or slashdot. Christian Faulhammer’s article from the 02.02 on pro-linux was a welcome exception. The proles among the proles (ubuntu and fedora users) are only seeing the bad stuff happening and ridicule the proles.
It also scares would-be-proles and turns them into opensuse users!
So it would help a lot if positive news would be posted on linux news sites. KDE 4.0.1 is in the tree? Great! The world needs to know that! Glibc 2.7 stable? Latest gcc unmasked, and testers asked for comment? unstable-users again saved from a security hole? Portage with some new feature? baselayout cleanups? Such stuff needs to be posted! Or not. Maybe releasing lots and lots of ‘testing stages’ would help. At least it would show the world and the proles that there is something happening besides adding ebuilds to the tree. Even if the stages are broken - that is why they are ‘testing’ stages.
Another thing is a felt disconnection. There aren’t many devs in the forums - and even less on the gentoo-user ml. Ok, gentoo-user is mostly doing fine (imho a lot better than the forums, almost no noise and much better answers), but still, a dev clarifying something or giving some tips&hints would be great.
Your explaining of GLEP 55 was great and well received (by me), I would like to see stuff like that more often - not only on planet.g.o but also on the ml and forums. Just a posting/thread about ‘we devs discussed this week glep XY, this are the points in laymen terms, if you want to comment on it, you can do it here and there’ ‘ or ‘what does it mean for the proles’, ‘would you like to see feature X or less wait for package BAR’ polls, or even’team xy needs more testers’ and other calls for help! Releng needs/wants some processing power to built&test stages and livecd images? Why not ask in the forums/ml? A lot gentoo users have highly overpowered machines. AT’ing might not for everybody. But the same person that is ‘too dumb’ to become a dev or AT might be smart&willing enough to built a bunch of livecds with half a douzend different settings and test-boot them. Make a ‘hall of fame’ for the most contributing users to give them a warm feeling. The final stages/images would be built by releng. But if some testing can be made by users and there is a simple way to reward them (like that ‘hall of fame’ - a little note somewhere) to give some fuzzy feelings, why not?
(as a side note, stuff like ‘we don’t care about users, this is made by devs for devs’ doesn’t really help. Yes we proles are mostly parasites. But for 10 parasites turned away by stuff like that, one contributing prole is lost as well…to pull some numbers out of my homungous behind - and it does make the gentoo devs look like a bunch of elitarian pricks)
And my third point is bugzilla. First, searching bugzilla is a pain in the ass. Whenever I have a problem, bugzilla is the first stop. And the interface really sucks. I hate the ‘zarro bugs found’ message, especially when I know that there are bugs that should have been listed. ALL helps - if it does not drown you in hits. Reporting bugs is almost worse. KDM does not show KDE 3.5 anymore, where to file? What to file? However the real problem with bugzilla is still the ‘friendly tone’ there. It got better, but is is far from perfect. It is bad enough when you are stuck somewhere, you don’t need a ‘closed as invalid’ without a real explaination in that situation - especially if a simple ‘closed as invalid because something with your setup is wrong, you need to check this and that, if the problem does not go away, ask here’ would get the user out of his problem (and I am not talking about the times when valid bugs are closed, because someone does not understand the bug …) The worst thing are dev-flamewars and power tripping in the comment sections (like with pci.ids and zlib useflag). Things like that should not happen.
February 23rd, 2008 at 7:04 am
Using Volker as a spring-board:
I’ll second the idea that positive news could be good. I think for that to happen there needs to be bigger accomplishments than “Random developer tweaks package to give himself an early birthday present.” I think it would require setting bigger distro wide goals and accomplishing them. Based on one of the other blogs, it sounds like that’s already been discovered though along with some other problems.
Don’t get me wrong, I like like to read flame eyes blog when he talking about things like his hunt for duplicate externals. I used to maintain a link-loader and that is interesting to me. It’s just not the sort of thing that will make it on slashdot or (name your favorite web site where the linux masses waste time). I don’t really care if gentoo gets good press really, but its probably good for gentoo to show that a source based distro can result in more than flame wars and spin off distros. It might even draw some more folks with time to be developers. That could address another known problem.
Also, I’m not as interested in developers gracing the masses with appearances in forums and mailing lists and such. It’s not that I’m against it. I just don’t read all that myself. I’d love it if using gentoo only required bugzilla for finding out what went bang and reading the newsletter or an rss feed for finding out what’s new or what could go bang. I’d still probably read some of the blogs since some of them are interesting and it recreational for me. More access to devs for when you try to contribute something but hear nothing back might be nice, but dev time is limited and there are trade offs there. The quick turn around on kernel exploits is worth a lot. Far more than, oh say, an init script for a game noone plays. Besides, if the devs are too busy to talk to human relations, they won’t have time hang out with the proles.
A hall of fame feels a little too childish too me. It has a game show feel to it, but if it would motivate enough people, fine. Make it optional though. Some people might rather hide.
Edd
February 23rd, 2008 at 10:39 am
To me, this is really simple. Gentoo _must_ have a stable platform for people to build against. There will _never_ be critical mass behind Gentoo until someone, somewhere, somehow builds a version that someone like an Oracle(1) can validate against. Until then, it is just a hobbyist’s distro. It simply can not be used for serious production work.
Don’t get me wrong. I use it all the time. I have 6 boxes running at home. But I would never recommend it for a project at work, which is notable because I am soon to be in a position to suggest that we move from Solaris to Linux for a major, worldwide application. Frankly, this is the only reason that Ubuntu has done what it’s done so quickly.
That being said, I even have a suggestion for how to do it. It would seem basically straightforward to create a “LTS” profile, like any other. It would simply have a package.mask filed with things like “<=sys-libs/glibc”. In this way, it would be simple to lock down a particular set of libraries and binaries to a supportable subset that could be validated against for a major application, Scalix or MySQL or such.
The problem here is that someone has to commit to back-porting security updates to this imaginary profile. On Gentoo, this doesn’t happen. All patches are applied to the current release, or a new release is made. This is exactly why I say you can’t use it for production. Open source is notorious for changing things on new releases (and that’s fine), but you can’t have things changing out from underneath you for the sake of security patches. This sort of thing is exactly what the SuSE’s and RedHat’s of the world are doing, and that’s why they are supported platforms for large ISV’s.
Am I crazy? Is this already happening and I just don’t know it? Should I actually be using Sabyon or something?
(1) http://download-uk.oracle.com/docs/cd/B25329_01/doc/install.102/b25144/toc.htm#BABGGAJA
February 23rd, 2008 at 10:41 am
Er… “<=sys-libs/glibc-2.6.1″, but you get what I was saying.
February 23rd, 2008 at 11:39 am
The following conversation is paraphrased and possibly never happened:
“I’ll have to talk to ___ to get the details.”
(week response)
“____ isn’t in right now, he’s on a business trip in orlando, he’ll be back in a week.”
“Ok, not really ____’s responsibility but he thinks he know someone that knows the code”
(several days response)
“Uh, that devs been retired for a bit now”
(couple days response)
“Yeah, I’ve been retired for awhile now and have just been informed”…
A > F > C > D and back to A.
(two months later)
“Ok it’s still not fixed yet. We got someone doing it but its an incomplete solution.”
“This kind of discussion belongs in the mailing list.”
Mailing List = puuuba
February 23rd, 2008 at 12:00 pm
Ah. _Perfect_ example “things changing out from underneath you” today as I updated my MythTV box:
http://www.gentoo.org/doc/en/apache-upgrading.xml
February 23rd, 2008 at 12:17 pm
emm, David Krider, why do you try to turn gentoo into RHEL?
There are many distro specialized for the ‘corporate workload’ why turn gentoo into another one? That ’stable plattform’ stuff would take away a lot of the things that make gentoo unique and beloved (remember, once gentoo was among the top 5 distros and at that point it was much more unstable than today).