11:55:43 #startmeeting 20140806 Flock T9:343 2:00pm Workstation 11:55:43 Meeting started Wed Aug 6 11:55:43 2014 UTC. The chair is yn1v. Information about MeetBot at http://wiki.debian.org/MeetBot. 11:55:43 Useful Commands: #action #agreed #halp #info #idea #link #topic. 11:56:42 #topic Fedora Work Station - Goals, Philosophy and Future by Christian F.K. Schaller 11:58:13 #info room for 25 people ant it has about 50. 11:59:36 More people is still coming 11:59:51 Fedora Workstation for the win! Can't wait to Fedora 21 :) 12:02:20 im going to try to live transcribe in here 12:02:31 * yn1v is relaying questions and comments 12:02:38 https://www.youtube.com/watch?v=LYtJZBgOrKw 12:03:07 Christian says he has been with RH for 3 years... manages Desktop team, installer team, and Fedora engineering team 12:03:14 why are we doing this 12:03:21 the desktop is not dying - 12:03:27 a lot has changed in marketplace 12:03:35 software consumption use cases are moving to phones and tablets 12:03:45 305 million units means you're not dying 12:03:48 thats a lot of potential users 12:03:57 the desktop is vibrant 12:04:03 the core areas of the desktop are slightly changing 12:04:09 less of consumption platform, more of a creation platform now 12:04:20 popular meme last few years - "this is the year of the linux desktop" 12:04:25 i think it actually is this time :) 12:04:36 chrome book 35% of all commercial laptop sales this year 12:04:41 steam now has over 600 games for linux 12:04:51 maybe this isnt what people were expecting 10 years ago but 12:05:01 does a console count as a desktop? steambox? 12:05:07 it creats a possible way for us to enter the market 12:05:10 same thing with chromebooks 12:05:17 it's not exactly gnome / kde / xfce what we expected but 12:05:26 as far as i can tell they are still doing very well esp in north america 12:05:43 so that's - combination of the fact there are still millions of users, makes me think it's worthwhile to do fedora workstation 12:05:48 so, philosophy 12:06:03 when we decided to do the workstation and fedora.next 12:06:17 needs to be overarching ideas on what we're trying to achieve here and how to achieve it 12:06:23 ive been stealing a few things from various people here 12:06:32 "We will not try to sell you. We will respect your privacy" 12:06:37 (stolen from matthew garrett?) 12:06:46 our opportunity - we're the only desktop that doesn't try to sell you 12:06:52 eg windows - trying to sell you to people 12:06:55 same with android 12:07:03 you're the product that they are selling to people to sell other products to you 12:07:15 fedora workstatino will not try to sell you something or sell your personal info to someone 12:07:25 and for the same reason - we are the desktop that respects your privacy 12:07:38 "double down on openness" 12:07:40 another principle 12:07:59 we shold make sure the platform is open, applications for it are open, provide... 12:08:00 the thing is 12:08:18 we're not in the business of running a store to compare with game studios or business platforms 12:08:34 a unique position in that we can say - we are a platform, anyone who comes to us can compete on equal terms 12:08:39 we're not trying to sell products to users 12:08:51 how we position ourselves with the workstation and fedora overall - 12:09:05 there's also... things that play particularly into workstation achieved by having 3 product model 12:09:13 "One product brings standards starting with appdata" 12:09:20 at this point - it's my opinion you can disagree - 12:09:33 fedora has been a loose collection of things... unified point has been things being packaged for Fedora 12:09:43 when we build these 3 products, we have a chance to define standards for the products 12:09:50 if you want to do something for Fedora, there's a way to do that 12:10:02 a lot over the past year, people asking - i want to port my app to linux or fedora - i never had a good answer 12:10:14 there's a million desktop, 500 kits, the way things work change basd on time of day 12:10:21 any integration point, it's hard to figure out what it's supposed to be 12:10:29 by having a clear product - workstation - we can start creating standards 12:10:43 if you're doing an app, and want to bring it to fedora - you can easily know that, you need to product appdata for your app 12:10:50 that's the standard way to provide metadata for your application 12:10:54 so that's what you need to provide 12:11:11 at the same point... for some reason, hey the standards... if i want to write something for fedora - i have to have something to parse and read that data 12:11:23 by having a clear open standard, makes it easier for exdternal and internal people to start doing things 12:11:30 there's no question about how it has to be done 12:11:34 "one product brings cohesion" 12:11:42 it helps us create standards, people can perform to 12:12:04 means, uip to this point, if you are an app developer - 12:12:37 every application - looks different, feels diferent, does things different before. with one product, then you can use a lot of toolkits and dtools to make the app, but have a more clear idea of what it needs to look like and how it needs to behave to fit in 12:12:44 that's a value having a fedora workstation brings us 12:13:03 so that's the high level position 12:13:11 value proposition about having a product like fedora workstation 12:13:19 for the first time, a clear idea of overall fedora 12:13:23 "future" 12:13:31 so, as we go forward, what does all this mean 12:13:43 doing a workstation or fedora server or anything else 12:13:48 or just providing standards... 12:13:53 can try to set technical direction 12:14:15 one thing i've been pushing hard inside of red hat - i think that traditional development model at red hat has been.... fedora as a test bed 12:14:18 for rhel 12:14:43 we need to set development goals from perspective of our products 12:15:02 and of cousre the good thing is that - for ppl who work at rh or community - being developers - we're not just memebers of fedora or rhel community 12:15:06 we're members of upstream communities 12:15:16 #chair mizmo 12:15:16 Current chairs: mizmo yn1v 12:15:18 if we need feature for workstation or server - we know how to make those features in a way that's acceptable to upstream 12:15:33 so that's - i guess - my approach as to how to do the workjstation 12:15:39 it's about what frame of mind we have in terms of approaching this 12:15:51 and this - what we need - and how do we align resources inside red hat and in the community to make sure we reach those goals 12:16:05 so some things we are definitely pushing for, going forward, in workstation 12:16:08 #1 thing is containers 12:16:11 "containers containers containers" 12:16:23 there's a big effort going on in general fedora land around docker 12:16:42 i think in fedora community there's a lot of diff opinions about doing app bundles and containizerized apps instead of sticking w the current model 12:16:48 but goes back to double down on open philosophy 12:16:56 containers provide us a chance to try to define something that can be used across distros 12:16:59 so we create a bigge rshared platform 12:17:05 as opposed to eveyrone being in their own (silo?) 12:17:23 classic comic strip - standards, try to fix it - 12:17:32 containers can also provide security benefits 12:17:38 also makes it a lot easier for vendors to support us 12:17:43 bc our platform will be less of a moving target 12:17:50 thats a lot of work happening around doing containers 12:17:58 we're been pushing hard on pre-reqs to do containers 12:18:07 talk about wayland - that is one core piece (mclasen's talk) 12:18:22 containers in X.... something there - they all share the same X context... so 12:18:29 last week, running a demo app to demonstrate that 12:18:38 special X extension custom made for... stealing your password (?) 12:18:42 so... we can fix that 12:18:50 big push we've been dfoing to try to get that to a useful level 12:18:53 we're getting close 12:19:13 we might not be there 100% but will be functional, hopefully have finalized later (?) 12:19:31 our users shouldn't need to know.... the more they are aware of this, the more we have failed in making this a seamless transition 12:19:40 another thing we've been working on in terms of containers 12:19:49 trying to define search (?)for them 12:19:57 containers in fedora based on containers in rhel, based on kernel 12:20:06 but desktop applications - not really relevant - a lot less relevant for apps 12:20:30 running app, want to be able to have access to a shared calendar for desktop, running inside a diff container or outside containers... acces to your webcam or microphone from inside a container... 12:20:41 so we need a lot more granular control - clear defined way these containers to reach out beyond the container 12:20:57 lot of stuff hapening around docker, definitely looking at docker to find out what peices we can really use 12:21:04 a lot of the tooling around docker specific to server side apps 12:21:13 at the moment primary interface for docker cont is ethernet/network port 12:21:24 what we are seeing - sql server on desktop - will just use standard container 12:21:33 if you want to run firefox in a container - that probably wont be a docker container per se 12:21:40 i mean we will share what we can we docker containers but 12:21:56 we have so many extrea layers on top of what docker providers, ppl on docker trademark might not want to call it docker container at that point 12:22:17 q: how do you standardize container, what services do you provide to it? (sounds like stephen tweedie) 12:22:45 for firefox - i need access to my file store - do you actually have a new API for firefox to call against this? or do you let any app build out - you lose commonality of file selection dialog. is this still being developed / discussed, is there a plan? 12:22:53 A: our plan at the moment for firefox, provide an API... 12:23:06 dbus... more public? you can have data transfer over dbus high speed and high volume 12:23:09 i think that's the current plan 12:23:17 if we want ot basically, you have to use this dbus api, and that will have... 12:23:32 Q: (st) in the container or outside the container, is the file chooser? is it talking to a file selection dialog? 12:23:48 A: (mclasen?) higher level api. app will ask i want to open a file, and the desktop env outside of the container will provide the UI for that 12:23:56 (back to talk) 12:24:01 christian: another core peice for desktop, 12:24:09 docker container traditionally. .. whole os to run in container? for desktop, 12:24:22 you dont necessarily want to have 500 of operating system for desktop in each container 12:24:30 we want to create standardized API of runtimes that containers can depend on 12:24:35 a nd call from inside the container 12:24:38 that's another piece of the puzzle 12:24:57 alex larsson, quite active with docker effort inside red hat 12:25:03 trying to do intiail writeup of these policise we will need 12:25:22 figure out where it makes sense to pull stuff from docker, or push stuff to docker, or do stuf fon the side that doesn't necessarily fit into docker bc so specific to desktop use case 12:25:25 "Grow the platform" 12:25:28 we also need to grow the platform 12:25:31 many ways of doing that 12:25:47 and of course - start pulling in more users that have shied away from fedora for various reasons 12:25:52 growing platform - very important 12:26:03 make sure fedora becomes natural stopping point for next gen of app developers 12:26:20 bc i think more often new efforts - open stack or docker or steamos and all these things 12:26:24 fedora has not been the platform of choice 12:26:41 we hope that we build the platform, make it better as a workstation for development, make sure that a few years down the line, the next big thing may pick fedora 12:26:47 or intial deployment platform will be fedora 12:26:50 rather than have to port later on 12:27:06 q: what steps are you planning to take publicity / grow platform? (sgallagh?) 12:27:21 (yes) 12:27:21 a: we haven't.... that push yet... esp me... we also blogging? build general community interest 12:27:35 we need to engage fedora marketing gruop and figure out how we can market the things we're trying to bring into workstation 12:27:36 to engage users 12:27:52 other thing i'm trying to look at - what we can achieve integration where we dont - at least fedora/ red hat doesn't (?) their own platform 12:28:00 certain apps targeting fedora... certain apps targeting rhel workstation 12:28:08 getting them to run in either env isn't necessarily an easy task 12:28:26 and of course, a lot big part of the reason for that in fedora... system wide rhel... 12:28:39 my hope is that that combination of apis we'll define for the runtimes and in containers 12:28:58 and also lookinga t waht we do - things we can do - from (?) to make it... when they compile they're going to have a bigger chance of... ? 12:29:26 i think that will be beneficial bc for rhel right, we issue, where their code didn't work in rhel 6, starting to use new APIs from glibc bc they were available in fedora 12:29:28 and vice-vesra 12:29:48 for instance - certified to run on rhel, but we have to make sure fedora users can run th app too, as we go forward and open up the potential userbase for fedora workstation 12:29:55 for instance ppl doing graphical design is a natural target beyond developers 12:30:02 we know that steam os... influence... 12:30:23 as far as... game developers... standard tools... we shold use that, see if we can make sure we have things in palce to mak sure that also works fine 12:30:29 (i guess he is talking about games not graphical design) 12:30:33 efforts like... 12:30:41 trying to just... belnder - more useful tool for game development 12:30:42 blender 12:30:52 "an actual ABI" 12:31:06 going back to containers... platform... we want to find a way to have an actual ABI that is the real thing 12:31:13 i mean, i guess we can in fedora, has been, 12:31:16 ... 12:31:21 hi langdon ? 12:31:32 whoever is smoking christian out please stop lol 12:31:48 inside fedora... a place for ABI... try to do here, define exdtended ABI for containers 12:32:03 we will try our best to make sure, it's going to stay compatible and available fo ra long time going forward 12:32:05 at the same time 12:32:15 provide platform... replacing apcakges... runtimes... 12:32:28 appication be sure the runtime version will work... and always at least for a long time period 12:32:38 but if you want access to new feature you need to port application to a new evgersion of the runtime 12:32:44 this is a crucial piece to making the platform work 12:32:54 we want to be less of a moving target, but we do not want to hold development back 12:32:58 so we'll have abi versions 12:33:04 q: langdon - will that be delivered by containers? 12:33:25 a: yes... define abis your container app will use... combo of dbus apis and desktop runtime... that's the hard thing - then the applications should be able to run against this for next 10 years 12:33:41 q: (stephen tweedie) back end changes to qt toolkit on abi so could run qt apps in container? 12:33:49 a: exepct to see... given set of runtimes defined 12:34:19 that... some cases might... have diff features... so... maybe you'll see we'll nottry to joint gtk/qt runtime, instead, have separate gtk and qt runtimes (?) 12:34:30 on fedora level, hoping to get containers runtimes could be more easily shared across distributions 12:34:37 "Become the platform for creators" 12:34:49 for me, i see it as the first step 12:35:06 bc as ive mentioned... i can see that, a lot of use cases are moving to tablets or phones in terms of viewing or viewing consuming media, or reading 12:35:19 people who create stuff... make art, make music, make code - i dont think laptop or desktop will go away for these people 12:35:24 i think it'll be their primary tool for a long time 12:35:29 as we go forward want o get developers on baord 12:35:37 as we go forward, get these different creative groups on board 12:35:44 available to build on the strengths we have 12:35:49 rhel is only linux version certified for maya 12:35:55 we want to extend that to have some support for fedora too 12:36:08 so people in fedora can run maya on their fedora workstation - greatly important app on the market 12:36:10 so.. 12:36:12 "challenges" 12:36:21 theres a lot of things we have to do to be successful here 12:36:26 marketing... that's deifnitely one item in the process 12:36:32 a lot of technical things we need to do relatively quickly 12:36:44 we cannot spend 10 years deinfing container tech / standards... train will have left station 12:36:50 "We have a lot of gruond to cover" 12:37:08 everyone's been involved in open source a long time.... continued situation of resource process.. bc in terms... we always feel like... amillion things that need to be done 12:37:18 so.... we really need to try to figure out better ways of working with our upstreams... 12:37:27 "big change for the fedora community" 12:37:35 fedora.next an d products a big change for fedora commnunity 12:37:46 i mean, discussions about these products and what they are trying to do and how they work, for a long time 12:38:04 i'm expecting as we move forwards into 3p product model, there will be a period for everyone involved, figuring out who does what, who decides, etc 12:38:11 "the interwebs" 12:38:12 big challenge 12:38:17 challenge for desktop itself 12:38:27 desktop apps... are becoming ... new desktop apps are becoming rarer and rarer 12:38:35 we have a lot of old projects... softwar ethat will remain crucial for all time 12:38:40 blender, gimp, inkscape... all these tools 12:38:43 im sure they'll stay around 12:38:49 but a lot of new things happening as web apps 12:38:56 a g reat opp for us... we can run web apps the same as anybody else 12:39:08 someone writes something for web we can support it, but then our platform becomes less of a target 12:39:25 i was looking at... red hat a couple of weeks ago... someone them, never open any app other than the brwoser 12:39:29 check email in browser 12:39:33 job involves 2-3 web apps 12:39:48 if all we need is webapps... 12:39:55 sort of makes desktop feel.... a broswer launcher 12:40:16 i think with workstation we're trying to make web apps feel more like they belong to desktop than something you just launch in a browser 12:40:26 theres no version of tweetdeck for linux - 12:40:33 i had never heard of it, knew it was related to twitter 12:40:43 it's literally tweetdeck.com or something - problem was not availability but discoverability 12:40:47 (it would work on linux) 12:40:49 easy request to fill 12:40:56 all we need to do is makesure we offer access to this web app 12:41:01 "usability vs ideology" 12:41:09 i think we'll have discussions in fedora comm about this 12:41:15 a lot of cases where we talk about usability vs ideology 12:41:21 tweetdeck is a good example 12:41:36 if we had in list of software center apps the user could have easily found it 12:41:44 but what is your view of web app / web page - not a free software app 12:41:58 i think that will be...a n ongoing challenge for us to go forward. navigate / find compromises to deal with tehse things 12:42:08 satisfy and make things easy for users, but make sure we dont lose our core values in the process 12:42:12 "Questions" 12:42:25 12:42:32 happy to take any q's or comments 12:43:09 q: (langdon) publicity side - i think aspect of targeting particular niche user groups is very helpful to publicity side. yeh technicall speaking it's a smaller overal body count than all computer users ever - but it would help to focus on targeted user gruop 12:43:41 yn1v, i have one comment if you could bring it up - basically we targeted creatives, designers, film makers, musicians at sxsw ~3 years ago and got a lot of interest 12:43:51 once we get critical mass of happy users from those groups 12:44:06 okey... I will relay 12:44:16 if the individual developers know the strategy - they make little decisions - ppl don't recognize devs make little decisions all the time that have a huge impact, clearly defined mission end up with better technical product 12:44:42 q: (adamw) what is the plan for service integration for goa - what is going to be built next, highest priority? or next api 12:44:58 a: dont have big roadmap but next hting we are discussing - better integration of (Calendar?) 12:45:08 currently there is some support, go into documents thing can print to... but that's one thing 12:45:20 also trying to figure out ways to intergrate to software development syustems out there 12:45:37 yn1v, i have another question if any thoughts about tighter integration with FLOSS app collab platforms like kolab and owncloud 12:45:54 a: (owen) goa - accounts is a tool kit 12:46:12 (christian) one thing to look at, talked about with spot - a way to integrate with fedora account system in goa, are there smart things we can do with that 12:46:26 yn1v, part of my q though - promiting those floss platforms while at the same time providing usability to the mainstream proprietary ones maybe 12:46:58 idea of trying to build desktop around apps - at the moment i think that's at least one big item on agenda is containers will help with some security issues 12:47:30 personally from time to time - still need to run, pdf viewer, less and less every year but some time document requires to run adobe pdf reader, would prefer to contain it bc of security issues 12:48:00 what about integration with, ensuring secure web pages by default, verification of certs through entire stack - i dont believe ... doe proper cert checking or verification lists or stuff like that 12:48:03 (who is asking the q?) 12:48:05 ssl services, etc 12:48:22 (owen) certainly it does... (???) could do more 12:48:40 (Christian) feel free to bring it up on fedora desktop list and we can take that discussion and figure out things we can do 12:48:56 a big part of success of fedora workstation or other products - not just about the current wg and what they want to do, but what everyone else can bring to the table 12:49:05 i am mroe than happy - if you have insights about security -0 would love to hear about it 12:49:09 (I think Peter was the person asking.) 12:49:37 thanks masta 12:50:27 jwb (?): not sure if desktop itself will take google services and figure out a way for owncloud to compete with it 12:50:35 but promiting it could happen 12:51:16 adamw: what i'd like to have - allow the effort servicse running - prob i have with it now - my mail is hosted on the same server... same domain as owncloud same identity - no way i can tell gnome that this identity on this server provides these servies - mail, owncloud, etc 12:51:27 i have to do this manually even though bascked by same freeipa installation 12:51:36 all these services kerberos backed 12:51:56 langdon: nice if there was identity mgmt settings - this user is attached to this service, smae user attached to this ervice, could change passwords at same time 12:52:02 yn1v, you did a good job bringing up my points thank you :) 12:52:12 jwb: i dont want to stop you - we are short on time and you are engineering things 12:52:25 jwb - providing access to those floss ervices and apps, can only be done so far but it's important 12:52:51 yn1v, dont know if you can tell jwb, that's not the point i was making at all. just wanted to make an easy on ramp for the FLOSS platforms 12:53:06 yn1v, we can better preserve our values if we promote the floss apps and make them veyr easy to use 12:53:20 thank you every body 12:53:21 12:53:28 And thank you mizmo 12:53:32 :) 12:53:47 Yes thanks mizmo 12:53:51 im going to live transcribe the outreach program for women talk in 302 next 12:54:05 #endmeeting