14:35:14 <qubodup> #startmeeting
14:35:14 <zodbot> Meeting started Fri Aug  9 14:35:14 2013 UTC.  The chair is qubodup. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:35:14 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
14:35:15 <qubodup> maybe
14:35:27 <qubodup> qubodup> ... i was sponsored to come over here
14:35:27 <qubodup> 17:04 < qubodup> ...based on an email that was written years ago
14:35:30 <qubodup> 17:04 < qubodup> the email was called 'why I hate fedora'
14:35:33 <qubodup> 17:04 < qubodup> there were bullets of points that sucked and i looked over them
14:35:36 <qubodup> 17:04 < qubodup> we faced quite a lot of them
14:35:39 <qubodup> qubodup> we don't quite suck that hard any more but there is room for improvement
14:35:42 <qubodup> 17:04 < qubodup> this is a rant
14:35:44 <qubodup> 17:04 < qubodup> we already improved a lot
14:35:47 <qubodup> 17:05 < qubodup> i wouldn't be working for fedora if we didn't things right but we're not perfect
14:35:50 <qubodup> 17:05 < qubodup> don't be offended :)
14:35:52 <qubodup> 17:05 < qubodup> I (hope) I won't be too harsh
14:35:55 <qubodup> let's over the technical issues
14:36:03 <qubodup> fedora is known for these. we are not afraid to break, rewrite, start from scratsh
14:36:12 <qubodup> we're pretty good at that and I think we should do this more often
14:36:21 <qubodup> on the other hand we have old problems that we don't address
14:36:27 <qubodup> . one of the major problems: fedora is fat
14:36:37 <qubodup> many people say it doesn't count, disk space is cheap
14:36:44 <qubodup> but I think with mobile and cloud it's important again
14:36:53 <qubodup> the problem is not that we're fat but we're getting more fat with every release
14:37:10 <qubodup> mainly - of course new softare, that's not a problem - but our default install gets bigger
14:37:18 <qubodup> even though we don't change the package set
14:37:42 <qubodup> if I have the same fedora of different versions, newer ones don't fit in the space of older ones
14:37:50 <qubodup> one of the reasons: more fonts.
14:38:06 <randomuser> *localization
14:38:08 <qubodup> they are really huge, different regions means more big fonts
14:38:30 <qubodup> our minimal install is not really minimal, instead it's ~700 MB ~250 packages
14:38:47 <qubodup> debian min install has 500MB and much less packages
14:38:50 <qubodup> (core)
14:39:08 <qubodup> ~90 packages
14:39:19 <qubodup> 50 running processes, while we have 68
14:39:31 <qubodup> this brings us to the next problem: we love daemons
14:39:44 <qubodup> have a problem? write a daemon
14:40:05 <qubodup> you don't need a daemon to automatically set the time, so I don't understand why we have one timezone daemon running all the time.
14:40:30 <qubodup> I think we don't need so many daemons, if we do write one, then please write c/c++, something efficient but not python
14:40:51 <qubodup> python is ok for config stuff, but not for daemons
14:41:04 <qubodup> another thing that became popular: dbus
14:41:12 <qubodup> I'm not completely against it, I think it's useful
14:41:32 <qubodup> but having core functionality depend on dbus that is not in the kernel
14:41:36 <qubodup> ... dbus can crash at any time
14:41:58 <qubodup> nobody really knows what happens when you have 5k users and everything is communicating over one dbus instance
14:42:12 <qubodup> so while dbus isn't in the kernel, it probably shouldn't be used for core functionality
14:42:51 <qubodup> when writing deamons, don't do[use?] (something like[?]) firewall-d.
14:42:56 <qubodup> just look at this syntax
14:43:12 <qubodup> you can't use any kind of bash completion for that
14:43:19 <qubodup> if you need to do something, please keep it simple and clean
14:43:34 <qubodup> firewall-d is one example of the features we have in the core. We don't really need it in core.
14:43:47 <qubodup> ip tables in core is fine in core, firewall-d is not
14:43:53 <qubodup> Anaconda.
14:43:59 <qubodup> some love it some hate it
14:44:01 <qubodup> I like it a lot
14:44:13 <qubodup> even though it needs a lot of configuration - something I didn't like in the beginning
14:44:24 <qubodup> this got better. nevertheless there are glitches
14:44:38 <qubodup> ...issues with button layouts...
14:44:47 <qubodup> * usability issues
14:45:06 <qubodup> we realized that many people don't find important buttons.
14:45:23 <qubodup> actually they have already reduced usability issues.
14:45:33 <qubodup> one issue i have is powerfull but non-intuitive partitioning
14:45:39 <qubodup> another is translations
14:45:50 <qubodup> the biggest issue is where we define the package sources[?]
14:46:36 <qubodup> ... you have a software selection thing where you set enviroment like kde and then options, which depend of what you pick in the next step
14:46:42 <qubodup> there are weird combinations possible
14:47:24 <qubodup> add minimal install, then add admin, which eventually includes graphical interfaces, but it's not clear what you get
14:47:35 <qubodup> sometimes it's random how dependencies are resolved
14:47:46 <qubodup> . It makes sense having a server with the design suite running on it. right?
14:48:16 <qubodup> there are two ways to install gnome, but no way to install kde and gnome at the same time
14:48:19 <qubodup> (or two desktops)
14:48:26 <qubodup> no way to install window managers
14:48:37 <qubodup> no way to install for example openbox, because it's not an environment
14:48:57 <qubodup> we had groups like 'web based internet' but we don't have this any more
14:49:12 <qubodup> whenever you change environment choices, the options get reset
14:49:18 <qubodup> GUI issues...
14:49:28 <qubodup> example time: basic-desktop is a group
14:49:37 <qubodup> it installs icons, themes...
14:49:49 <qubodup> initial setup (not sure this makes sense)
14:50:23 <qubodup> you get 6 different window managers
14:50:33 <qubodup> I think 'basic-desktop' should give you ONE very minimal window manager
14:50:48 <qubodup> now to core: does anybody know why we have plymouth in this?
14:51:05 <qubodup> even if there was? in that case it would be broken and we'd need to fix it.
14:51:15 <qubodup> then there's the sendmail discussion. I'm not sure you are all aware of it
14:51:33 <qubodup> ... wanted to remove it from core, but it got rejected by fasco[?]. it was 4v4 one undecided.
14:51:50 <mattdm> fasco = FESCO
14:51:51 <qubodup> "actually it's not in core any more, it's in standard now"
14:52:05 <mattdm> (also that's not exactly what happened.)
14:52:07 <qubodup> If that's fixed, that's great.
14:52:19 <nirik> fiaisco?
14:52:20 <nirik> :)
14:52:29 <qubodup> dial-up is included in all envs. who ever used this?
14:52:54 <qubodup> I worked at a telephone provider for a few years. i was one of the few who used isdn.
14:53:11 <qubodup> "isdn is an edge case, useful for skype"
14:53:25 <qubodup> in germany you can't even buy isdn, unless you're a special case.
14:53:34 <qubodup> You usually get voip
14:53:42 <qubodup> I don't know if we need any of this modem stuff
14:53:48 <qubodup> of these modem packages
14:54:00 <qubodup> "there's a case where they need that..."
14:54:13 <qubodup> "if you live in that world, keep it installed for yourself[?]"
14:54:26 <qubodup> Ok, now the developer environment
14:54:40 <qubodup> you hve java development, php is optional
14:54:47 <qubodup> we have font-design in the design suite
14:54:56 <qubodup> ... I don't think a designer is a java developer at the same time
14:55:05 <qubodup> in all environments but minimum we have fonts
14:55:18 <qubodup> This is easy to clean up (joking) it's a lot of work actually
14:55:26 <qubodup> for a start we need to reduce what 'core' is
14:55:35 <qubodup> for example debian doesn't even ship openssh in their minimal install
14:55:45 <qubodup> they have something that is similar to ...[?]
14:56:00 <qubodup> ... make it as small as possible
14:56:03 <qubodup> use reasonable extends
14:56:08 <qubodup> clean up everything else you find
14:56:16 <qubodup> i'd like to see individual package selection in anaconda
14:56:19 <qubodup> "never gonna happen"
14:56:28 <qubodup> "but we're open to ideas. but not to package selection"
14:56:34 <qubodup> "*individual"
14:56:49 <qubodup> we still have the optional packages in coms[?] but anaconda doesn't use that
14:56:57 <qubodup> i like th old model where you have group webserver
14:56:59 <sgallagh> qubodup: "comps" aka comps.xml
14:57:02 <qubodup> and then you can extend that
14:57:28 <qubodup> currently... think of the new situation and window managers. you have to define one group for each WM even if it's only one package
14:57:37 <qubodup> "it needs more work but the old way is not the solution"
14:57:41 <qubodup> this is like windows2000
14:57:51 <qubodup> where you need to install games[?] on a server and then you need to remove it
14:58:04 <qubodup> "talk to mo[?] we need more ideas"
14:58:10 <qubodup> "we have 2 ideas and all UI people hate one of them"
14:58:28 <qubodup> "we should[?] have something based on functionality, not on packages"
14:59:07 <qubodup> "...if you don't like it, it's open source..."
14:59:12 <qubodup> I think we should agree on a same set of fonts
14:59:21 <qubodup> you can't support every random language of the world by default
14:59:37 <qubodup> we should have an installation medium that addresses that % of that population
14:59:50 <qubodup> but many don't want to install ~50MB of fonts
15:00:05 <qubodup> "some of that stuff is really unclear, like dictionaries for certain functionalities"
15:00:18 <qubodup> "if you remove it, does this mean the dict is incomplete or doesn't work?"
15:00:24 <qubodup> "it would be nice to have that clearer"
15:00:44 <qubodup> "we should have standard languages and have the option to pick other languages during intallation"
15:01:08 <qubodup> "gettext i hard"
15:01:59 <qubodup> packaging itself... we have huge packages. i'm not saying we should do it like debian
15:02:10 <qubodup> i think debian has this in ...[?] packages. that's insane
15:02:22 <qubodup> you shouldn't have monolithic packages, especially if it extens something
15:02:32 <qubodup> you will want foo-nautilus, to just extend nautilus
15:02:42 <qubodup> so most of the users' don't need to install more than they want
15:03:11 <qubodup> ..making dependencies shorter it's good to see that a lot of people are willing to work on this.
15:03:22 <qubodup> ... at the moment xfce doesn't have bluetooth interface
15:03:28 <qubodup> and you would fallback to the one of gnome
15:03:33 <qubodup> so you can fallback to the one of mate
15:03:40 <qubodup> which is the old gnome bluetooth packages
15:03:59 <qubodup> but this install mate-session, mate-panel... ~50MB, again dependency bloat
15:04:12 <qubodup> "istn't this solveable by optional deps?"
15:04:22 <qubodup> We have been begging for opt deps for years...
15:04:28 <qubodup> it's taking another 5 centuries
15:04:37 <qubodup> let's fix the issues not upstream but now
15:04:45 <qubodup> Any ABRT guys here?
15:04:55 <qubodup> No? OK, then I can bas it! (No I won't) :)
15:05:02 <qubodup> Anonymous bug reporting is uselesss
15:05:15 <qubodup> I know people like it but it only works if you respond on time
15:05:47 <qubodup> a freidn of mine works with anonymous reports. it works if you respond in 12h but do it later and they won't look it up again
15:06:06 <qubodup> so its' bad for code bugs which are upstream problems most of the time
15:06:16 <qubodup> ... as a packager it keeps me from doing my job ...
15:06:27 <qubodup> at the momen where I have listed 200 bugs
15:06:32 <qubodup> i get crazy asking for questions
15:06:43 <qubodup> ... then translations
15:06:53 <qubodup> they are not a problem for most parts of the OS
15:07:09 <qubodup> to fix this we use transifex which works for most things
15:07:30 <qubodup> ..filing something in bugzilla wil be closed as WONTFIX
15:07:44 <qubodup> if we upstream for something, it means we should upstream transifex as well
15:07:56 <qubodup> this means translations should use transifex and do it upstream
15:08:03 <qubodup> and have a real upstream translation community
15:08:10 <qubodup> i don't know if we can abuse the fedora doc team
15:08:34 <qubodup> ... germany is covered well but usually translated out of context
15:09:01 <qubodup> it looks like people don't test their translations and don't notice konjugatyion errrors
15:09:10 <qubodup> when you translate, please test in the context of usage
15:09:15 <qubodup> Now about community
15:09:19 <qubodup> I won't cover too much
15:09:25 <qubodup> I think it's where we improved mose
15:09:26 <qubodup> most
15:09:37 <qubodup> The division between contributors and users is a problem
15:09:40 <qubodup> mailing list vs forums
15:09:44 <qubodup> does anybody use the forums?
15:09:49 <qubodup> does anybody answer questions there?
15:09:56 <qubodup> "I ask questions there"
15:10:00 <qubodup> does anybody use askfedora?
15:10:09 <qubodup> oh its' more popular than the forum?
15:10:34 <qubodup> We have a german forum, which are enthusiasts but not "official" fedora members
15:10:56 <qubodup> "to me contributor means they contribute"
15:11:08 <qubodup> to me a contributor is everybody with a fedora account
15:11:18 <qubodup> one of the downside is we don't have end user doc in the wiki
15:11:46 <qubodup> ubuntu has every tool - how open gnome-terminal, how to become SU, they have beginner and advanced covered in the wiki
15:11:55 <qubodup> We should have one central point to access all this info
15:12:15 <qubodup> just in germany we have fedora forum, blog, wiki, run by different probelm, overlapping and contradictinng
15:12:19 <qubodup> this is a problem i think
15:12:20 <qubodup> next point
15:12:25 <qubodup> lawyers are a problem
15:12:48 <qubodup> in europ we have a nonprofit which we had to close because "Fedora" couldn't be used
15:12:51 <qubodup> trademark stuff
15:12:58 <qubodup> I understand redhat's protetion requirements
15:13:04 <qubodup> but when it comes to swag
15:13:13 <qubodup> we won't be able to make any decent swag this way
15:13:19 <qubodup> debian is great at that
15:13:36 <qubodup> for me the trademark issues..
15:13:45 <qubodup> we have achieved a lot but still the lawyers stuff is a pain
15:13:50 <qubodup> and it'ns not gonna be better at the moment
15:13:57 <qubodup> because there's a lot of stuff in flux
15:14:07 <qubodup> i don't have a solution for these points
15:14:14 <qubodup> i hope we will find time to discuss them here at flock
15:14:26 <qubodup> next: flock - flock is a community problem
15:14:40 <qubodup> we had a process for fudcons but we didn't use it for flock
15:14:48 <qubodup> i like the idea of a new idea and of a global event
15:14:57 <qubodup> but I know people in other regions are sad that they can't make it
15:15:00 <qubodup> it's expensive to get here
15:15:10 <qubodup> if you want to get here, most students and so rely on sponsoring
15:15:15 <qubodup> the process for that was intransparent
15:15:28 <qubodup> last time we had the team discussing should we sponsor that or this person
15:15:36 <qubodup> you could recommend people
15:15:58 <qubodup> we have two people from famsco, south america, who were not sponsored to get here
15:16:01 <qubodup> i think that's really sad
15:16:09 <qubodup> but it's probably not the right time to discsuss this at flock
15:16:15 <qubodup> let's address theo ther problems
15:16:26 <qubodup> last but not least
15:16:31 <qubodup> suckin is good
15:16:46 <qubodup> * let's addres the other problems
15:16:47 <qubodup> last point: sucking is good
15:17:19 <qubodup> we are not necessarily the most easy distro to use, I don't thing that's bad
15:17:26 <qubodup> we can make it more user friendly
15:17:34 <qubodup> but let's fix them on the right way, not go the ubuntu way
15:17:47 <qubodup> which seems to make it super easy for anybody to do anything.
15:18:02 <qubodup> and if you add proprietary drivers, you don't make things easier exept for one case
15:18:06 <qubodup> let's try to suck less
15:18:15 <qubodup> by the way, suckless.org has aimilar statment
15:18:41 <qubodup> saying that config files make newbies go away and not ask questions
15:18:53 <qubodup> let's work on issues here at flock
15:19:32 <qubodup> #stopmeeting
15:19:36 <nb> #fedora-flock-ectr103
15:19:36 <nb> #fedora-flock-ectr107
15:19:37 <nb> #fedora-flock-ectr109
15:19:38 <nb> #fedora-flock-ectr112
15:19:39 <nb> #fedora-flock-ectr114
15:19:41 <nb> #fedora-flock-ectr115
15:19:43 <nb> #fedora-flock-ectr120
15:19:45 <nb> #fedora-flock-ectr212
15:19:47 <nb> oops
15:19:52 <qubodup> #endmeeting