Hi All,
Since everyone else seems to be shy about kicking off some of the threads for the next steps, I'll get this one going and get it out of the way.
We need to settle on an underlying DE for the Workstation product. The two major DEs in the Linux space are GNOME and KDE. Fedora has spins for MATE, XFCE, and one other (I think). I've gathered that there's a lot of assumption, both in the broader community and within the WG, that Workstation will continue the Fedora trend and be based on GNOME. I would even venture to say that is a fairly sane assumption to make.
With that in mind, would the WG like to officially settle on using GNOME as the underlying DE for Workstation?
I will be perfectly honest and say I have no overwhelming preference here personally. My expertise extends to helping navigate through Fedora process thus far, so I'm not sure I'd make a huge impact from the technical side of things on whatever DE is picked.
josh
I would vote for Cinnamon as the default DE. If not Cinnamon, then KDE would be my next choice.
On Thu, Jan 30, 2014 at 3:47 PM, Josh Boyer jwboyer@fedoraproject.orgwrote:
Hi All,
Since everyone else seems to be shy about kicking off some of the threads for the next steps, I'll get this one going and get it out of the way.
We need to settle on an underlying DE for the Workstation product. The two major DEs in the Linux space are GNOME and KDE. Fedora has spins for MATE, XFCE, and one other (I think). I've gathered that there's a lot of assumption, both in the broader community and within the WG, that Workstation will continue the Fedora trend and be based on GNOME. I would even venture to say that is a fairly sane assumption to make.
With that in mind, would the WG like to officially settle on using GNOME as the underlying DE for Workstation?
I will be perfectly honest and say I have no overwhelming preference here personally. My expertise extends to helping navigate through Fedora process thus far, so I'm not sure I'd make a huge impact from the technical side of things on whatever DE is picked.
josh
desktop mailing list desktop@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/desktop
On Thu, 2014-01-30 at 15:47 -0500, Josh Boyer wrote:
Hi All,
Since everyone else seems to be shy about kicking off some of the threads for the next steps, I'll get this one going and get it out of the way.
We need to settle on an underlying DE for the Workstation product. The two major DEs in the Linux space are GNOME and KDE. Fedora has spins for MATE, XFCE, and one other (I think). I've gathered that there's a lot of assumption, both in the broader community and within the WG, that Workstation will continue the Fedora trend and be based on GNOME. I would even venture to say that is a fairly sane assumption to make.
With that in mind, would the WG like to officially settle on using GNOME as the underlying DE for Workstation?
I will be perfectly honest and say I have no overwhelming preference here personally. My expertise extends to helping navigate through Fedora process thus far, so I'm not sure I'd make a huge impact from the technical side of things on whatever DE is picked.
Given the tension between the definition of a "Workstation Product" and the multiple desktop spins that I've identified on devel@ - i.e. that a "Workstation product" built around a single desktop occupies the 'desktop space', without accounting for alternative desktops - do you definitely want to go ahead with the model where the WS product is specifically associated with a single desktop and makes no attempt to somehow 'include' alternative desktops, or is it worth considering possible approaches that somehow account for alternatives? I realize it might be quite late to do that, but it seemed worth asking the question.
On Thu, Jan 30, 2014 at 3:55 PM, Adam Williamson awilliam@redhat.com wrote:
On Thu, 2014-01-30 at 15:47 -0500, Josh Boyer wrote:
Hi All,
Since everyone else seems to be shy about kicking off some of the threads for the next steps, I'll get this one going and get it out of the way.
We need to settle on an underlying DE for the Workstation product. The two major DEs in the Linux space are GNOME and KDE. Fedora has spins for MATE, XFCE, and one other (I think). I've gathered that there's a lot of assumption, both in the broader community and within the WG, that Workstation will continue the Fedora trend and be based on GNOME. I would even venture to say that is a fairly sane assumption to make.
With that in mind, would the WG like to officially settle on using GNOME as the underlying DE for Workstation?
I will be perfectly honest and say I have no overwhelming preference here personally. My expertise extends to helping navigate through Fedora process thus far, so I'm not sure I'd make a huge impact from the technical side of things on whatever DE is picked.
Given the tension between the definition of a "Workstation Product" and the multiple desktop spins that I've identified on devel@ - i.e. that a "Workstation product" built around a single desktop occupies the 'desktop space', without accounting for alternative desktops - do you definitely want to go ahead with the model where the WS product is specifically associated with a single desktop and makes no attempt to somehow 'include' alternative desktops, or is it worth considering possible approaches that somehow account for alternatives? I realize it might be quite late to do that, but it seemed worth asking the question.
The PRD already has a section that speaks to working with various toolkits to make them inter-operate with each other. The idea behind that is so that applications from various toolkits have the same look-and-feel on the Workstation regardless of their primary DE/toolkit. While I realize that isn't specifically answering your question, it does at least speak to the fact that there isn't ONE TOOLKIT TO RULE THEM ALL.
As to whether we want Workstation to be a "pick your own DE" product, I personally don't feel that's a great way to start. Seems somewhat confusing. There's a lot of work to be done, and having a single underlying toolkit/platform to work from at least helps to focus on where the initial work goes. I also don't think it excludes the possibility of different Workstation DEs in the future. If we'd like to rephrase as "initial Workstation DE" to allow for that possibility, I'm OK with that.
josh
On Thu, Jan 30, 2014 at 3:54 PM, Lynn Dixon boodaddy@gmail.com wrote:
I would vote for Cinnamon as the default DE. If not Cinnamon, then KDE would be my next choice.
Thanks for replying!
The thread isn't intended to be a popularity vote. I'm sure everyone has their favorite DE that they'd like to see as the basis for Workstation. If you have compelling reasons that Cinnamon should be that underlying DE, I'd encourage you to bring them forward and present a well reasoned rationale. Those kinds of details are important in the consideration, not just personal preference.
I will admit that my initial thread didn't list the reasoning behind GNOME, so I'm somewhat guilty of that myself. I was coming from that stance that this has been debated numerous times within the Fedora project and I was hoping to avoid a rehash on the strong position GNOME occupies in our project. In short, they boil down to man power, upstream contribution, familiarity, and continued testing based on previous releases of Fedora. I'm certain that there are others in the WG that can expand on that if _really_ required, but I would like it if we could avoid duplicating years worth of discussion.
josh
On Thu, 2014-01-30 at 16:08 -0500, Josh Boyer wrote:
Given the tension between the definition of a "Workstation Product" and the multiple desktop spins that I've identified on devel@ - i.e. that a "Workstation product" built around a single desktop occupies the 'desktop space', without accounting for alternative desktops - do you definitely want to go ahead with the model where the WS product is specifically associated with a single desktop and makes no attempt to somehow 'include' alternative desktops, or is it worth considering possible approaches that somehow account for alternatives? I realize it might be quite late to do that, but it seemed worth asking the question.
The PRD already has a section that speaks to working with various toolkits to make them inter-operate with each other. The idea behind that is so that applications from various toolkits have the same look-and-feel on the Workstation regardless of their primary DE/toolkit. While I realize that isn't specifically answering your question, it does at least speak to the fact that there isn't ONE TOOLKIT TO RULE THEM ALL.
The bit of the PRD that seems significant is "The Workstation working group will define a set of packages that are considered required be installed in order for the system to qualify as a Fedora Workstation. Through policy users will be strongly advised against uninstalling any of these packages and there will also be no option in the graphical software installer to uninstall them."
I was kinda reading that as essentially mandating that the default desktop must be installed. It doesn't preclude installing another alongside it, but the effect seems to be to define a standard environment which is always going to be assumed and prioritized.
As to whether we want Workstation to be a "pick your own DE" product, I personally don't feel that's a great way to start. Seems somewhat confusing. There's a lot of work to be done, and having a single underlying toolkit/platform to work from at least helps to focus on where the initial work goes. I also don't think it excludes the possibility of different Workstation DEs in the future. If we'd like to rephrase as "initial Workstation DE" to allow for that possibility, I'm OK with that.
That wasn't necessarily what I was suggesting, it was more of an open suggestion than a specific implementation idea. There are probably approaches that don't involve the "Workstation product" per se being a choose-your-own-adventure, but somehow allow for the existence of alternatives. I think someone floated the idea of some kind of sub-product system already, for instance.
On Thu, Jan 30, 2014 at 4:19 PM, Adam Williamson awilliam@redhat.com wrote:
On Thu, 2014-01-30 at 16:08 -0500, Josh Boyer wrote:
Given the tension between the definition of a "Workstation Product" and the multiple desktop spins that I've identified on devel@ - i.e. that a "Workstation product" built around a single desktop occupies the 'desktop space', without accounting for alternative desktops - do you definitely want to go ahead with the model where the WS product is specifically associated with a single desktop and makes no attempt to somehow 'include' alternative desktops, or is it worth considering possible approaches that somehow account for alternatives? I realize it might be quite late to do that, but it seemed worth asking the question.
The PRD already has a section that speaks to working with various toolkits to make them inter-operate with each other. The idea behind that is so that applications from various toolkits have the same look-and-feel on the Workstation regardless of their primary DE/toolkit. While I realize that isn't specifically answering your question, it does at least speak to the fact that there isn't ONE TOOLKIT TO RULE THEM ALL.
The bit of the PRD that seems significant is "The Workstation working group will define a set of packages that are considered required be installed in order for the system to qualify as a Fedora Workstation. Through policy users will be strongly advised against uninstalling any of these packages and there will also be no option in the graphical software installer to uninstall them."
I was kinda reading that as essentially mandating that the default desktop must be installed. It doesn't preclude installing another alongside it, but the effect seems to be to define a standard environment which is always going to be assumed and prioritized.
I was speaking more to these two sections:
"
Work towards standardizing and unifying the Linux desktop space
We want to use and develop technologies that can be widely shared with the rest of the community and we want to allow developers to use the tools they prefer for their application development yet make them all feel like a natural fit into our integrated desktop experience. This would include items like theming and making sure we offer a consistent accessibility story across different development toolkits. The product will reach out and collaborate with the Fedora Design team and other relevant groups on such items.
Develop application guidelines and designs
The working group will develop guides and style recommendations for applications that target the workstation. These guidelines will be mandatory for the core apps that are specifically developed for the workstation, but 3rd party software developers will be encouraged to follow them too.
"
which reads more like what I was referring to. Collaboration with other applications and toolkits to make sure a wide variety of software works as expected on Workstation.
As to whether we want Workstation to be a "pick your own DE" product, I personally don't feel that's a great way to start. Seems somewhat confusing. There's a lot of work to be done, and having a single underlying toolkit/platform to work from at least helps to focus on where the initial work goes. I also don't think it excludes the possibility of different Workstation DEs in the future. If we'd like to rephrase as "initial Workstation DE" to allow for that possibility, I'm OK with that.
That wasn't necessarily what I was suggesting, it was more of an open suggestion than a specific implementation idea. There are probably approaches that don't involve the "Workstation product" per se being a choose-your-own-adventure, but somehow allow for the existence of alternatives. I think someone floated the idea of some kind of sub-product system already, for instance.
If people want alternatives to exist, there's no reason to somehow prohibit them from doing that. Nor do I think we'd want to even think about preventing them. I don't necessarily think those alternatives should be grouped or branded under Workstation though.
I feel like I'm not understanding what you're asking. Could you maybe try to elaborate a bit more with examples?
josh
On Thu, 2014-01-30 at 16:29 -0500, Josh Boyer wrote:
That wasn't necessarily what I was suggesting, it was more of an open suggestion than a specific implementation idea. There are probably approaches that don't involve the "Workstation product" per se being a choose-your-own-adventure, but somehow allow for the existence of alternatives. I think someone floated the idea of some kind of sub-product system already, for instance.
If people want alternatives to exist, there's no reason to somehow prohibit them from doing that. Nor do I think we'd want to even think about preventing them. I don't necessarily think those alternatives should be grouped or branded under Workstation though.
I feel like I'm not understanding what you're asking. Could you maybe try to elaborate a bit more with examples?
OK, let me take another shot :)
As I suggested on devel@, conceiving of Fedora in the 'three product design' - Fedora is the Workstation, Cloud and Server products - leaves the status of things that were previous Fedora but do not fit neatly into that 'product' definition somewhat up in the air. Just as notting put it, what exactly is the status of spins, especially alternative desktop spins?
Workstation WG could say you want to take a shot at resolving that problem somehow; I'm just asking if that's something you're interested in doing, or if you'd rather it be resolved through some other group/process.
On Thu, Jan 30, 2014 at 4:58 PM, Adam Williamson awilliam@redhat.com wrote:
On Thu, 2014-01-30 at 16:29 -0500, Josh Boyer wrote:
That wasn't necessarily what I was suggesting, it was more of an open suggestion than a specific implementation idea. There are probably approaches that don't involve the "Workstation product" per se being a choose-your-own-adventure, but somehow allow for the existence of alternatives. I think someone floated the idea of some kind of sub-product system already, for instance.
If people want alternatives to exist, there's no reason to somehow prohibit them from doing that. Nor do I think we'd want to even think about preventing them. I don't necessarily think those alternatives should be grouped or branded under Workstation though.
I feel like I'm not understanding what you're asking. Could you maybe try to elaborate a bit more with examples?
OK, let me take another shot :)
As I suggested on devel@, conceiving of Fedora in the 'three product design' - Fedora is the Workstation, Cloud and Server products - leaves the status of things that were previous Fedora but do not fit neatly into that 'product' definition somewhat up in the air. Just as notting put it, what exactly is the status of spins, especially alternative desktop spins?
In my mind the KDE spin is the only one that doesn't have a clear cut status in the new world. That's possibly because of some assumptions I've made that may or may not be accurate. Anyway.
Workstation WG could say you want to take a shot at resolving that problem somehow; I'm just asking if that's something you're interested in doing, or if you'd rather it be resolved through some other group/process.
Ah. To be honest, I think that's up to the FESCo/Board level. There are spins that extend beyond just choice of DE, and people could want to create different spins of Server, etc. So I personally wouldn't advocate for the Workstation WG to solve that issue. It's definitely something that needs to be worked out, I just think it needs to be done at a higher level.
josh
On 01/30/2014 09:55 PM, Adam Williamson wrote:
Given the tension between the definition of a "Workstation Product" and the multiple desktop spins that I've identified on devel@ - i.e. that a "Workstation product" built around a single desktop occupies the 'desktop space', without accounting for alternative desktops - do you definitely want to go ahead with the model where the WS product is specifically associated with a single desktop and makes no attempt to somehow 'include' alternative desktops, or is it worth considering possible approaches that somehow account for alternatives? I realize it might be quite late to do that, but it seemed worth asking the question.
I personally feel that a single default offering is a must, if Fedora is to be successful in the desktop market. We have been losing market share to Ubuntu that has one single default desktop product, and I think this is a lesson to learn from.
When I joined the Workstation WG, I did that to help build a successful product. To build a base system system that user can rely on; a base system that 3rd party vendors can reliably target with their software. Most other WG members I've talked to are also here to help build a single product.
I do not want to downplay the value of Spins and alternative offerings, but I personally do not want to spend my time developing them, and I'd rather see if they were developed elsewhere and the Workstation WG was limited to putting together one product.
On Thu, 30 Jan 2014 16:08:10 -0500 Josh Boyer jwboyer@fedoraproject.org wrote:
I also don't think it excludes the possibility of different Workstation DEs in the future
I haven't read the final prd. Will initial default product allow for User Foo yum erase Default Desktop, yum install DE of choice. Whether by make make install or groupinstall DE
___ Regards, Frank www.frankly3d.com
So the way I have seen this working in my mind so far is like this.
We install the default Workstation product, which uses the GNOME Shell. Included in the workstation default install there are also some other major libraries that is commonly used, Qt for instance comes to mind here as a obvious one. We try to provide what integration we can for any included library so that the user doesn't ideally have to think about technical details such as toolkit used by a developer. We also need to make sure that our accessibility stuff, like the high contrast icons, is equally supported across toolkits shipped.
Through the Software installer there will be other options available for install, like KDE. These will be installed alongside the default package set. We will not have a replace default option, just an add one, in order to ensure that the default package set can be targeted by 3rd parties.
Of course for power users there is still the command line tools and we are not going to do anything to try stop people from using those command line tools to change their system however they want. Only thing I can see us doing is that if we at some point is to maybe cancel automatic bug reporting through brt on systems that have seen manual surgery (not sure how we detect that), to avoid distorting our bug statistics with user caused issues, but that is an issue for later times in the case we end up having data indicating that it actually is an issue.
Jaroslav mentioned the idea of formulas and while not exactly the same we are doing here for desktops it is in my opinion along the same line of thinking. We might want to consider doing something along these lines for other things than desktops too, like for instance having some kind of profile packs targeting specific user groups, but I guess it might be something we want to not put on the agenda right away, to have a chance to ponder a bit more what those packs would be and how we present them for installation. My main concern is that if we create to many 'meta' packages the installer will start looking rather cluttered, but maybe the Fedora and Desktop design teams has some input regarding this. Of course one could argue that the whole point of the software installer is to make it so easy to find stuff that we don't need this, and that having these kind of 'group' installs is as accurate as shooting a sparrow with a cannon.
I also took the liberty of adding a Technical Speification page to the wiki: https://fedoraproject.org/wiki/Workstation/Technical_Specification
It is currently more of a skeleton than anything else and I don't think I am the right person to write it, but I know Matthias and others are interested in helping flesh it out and add details, and hopefully we can have a good discussion on this list about potential items to include, wording and scoping etc.
Christian
----- Original Message -----
From: "Adam Williamson" awilliam@redhat.com To: "Discussions about development for the Fedora desktop" desktop@lists.fedoraproject.org Sent: Thursday, January 30, 2014 9:55:46 PM Subject: Re: Underlying DE for the Workstation product
On Thu, 2014-01-30 at 15:47 -0500, Josh Boyer wrote:
Hi All,
Since everyone else seems to be shy about kicking off some of the threads for the next steps, I'll get this one going and get it out of the way.
We need to settle on an underlying DE for the Workstation product. The two major DEs in the Linux space are GNOME and KDE. Fedora has spins for MATE, XFCE, and one other (I think). I've gathered that there's a lot of assumption, both in the broader community and within the WG, that Workstation will continue the Fedora trend and be based on GNOME. I would even venture to say that is a fairly sane assumption to make.
With that in mind, would the WG like to officially settle on using GNOME as the underlying DE for Workstation?
I will be perfectly honest and say I have no overwhelming preference here personally. My expertise extends to helping navigate through Fedora process thus far, so I'm not sure I'd make a huge impact from the technical side of things on whatever DE is picked.
Given the tension between the definition of a "Workstation Product" and the multiple desktop spins that I've identified on devel@ - i.e. that a "Workstation product" built around a single desktop occupies the 'desktop space', without accounting for alternative desktops - do you definitely want to go ahead with the model where the WS product is specifically associated with a single desktop and makes no attempt to somehow 'include' alternative desktops, or is it worth considering possible approaches that somehow account for alternatives? I realize it might be quite late to do that, but it seemed worth asking the question. -- Adam Williamson Fedora QA Community Monkey IRC: adamw | Twitter: AdamW_Fedora | XMPP: adamw AT happyassassin . net http://www.happyassassin.net
-- desktop mailing list desktop@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/desktop
On 31 January 2014 10:47, Christian Schaller cschalle@redhat.com wrote:
Through the Software installer there will be other options available for install, like KDE. These will be installed alongside the default package set. We will not have a replace default option, just an add one, in order to ensure that the default package set can be targeted by 3rd parties.
This is an interesting one. Do we want an application installer (that's my view of what gnome-software is) advertising and enabling users to install something like a totally different desktop environment?
Rui
That is a good question and there is always the risk of 'overloading' the app with to many purposes. That said in the next version there will be support for installing fonts, codecs and ibus methods, so there is some precedence for installing things that is technically not an 'application'.
Hopefully some of the designers can chime in on this, I mean we could also considering putting this functionality somewhere else than the application installer if the designers thinks that is a better option.
Christian
----- Original Message -----
From: "Rui Tiago Cação Matos" tiagomatos@gmail.com To: "Discussions about development for the Fedora desktop" desktop@lists.fedoraproject.org Sent: Friday, January 31, 2014 11:00:09 AM Subject: Re: Underlying DE for the Workstation product
On 31 January 2014 10:47, Christian Schaller cschalle@redhat.com wrote:
Through the Software installer there will be other options available for install, like KDE. These will be installed alongside the default package set. We will not have a replace default option, just an add one, in order to ensure that the default package set can be targeted by 3rd parties.
This is an interesting one. Do we want an application installer (that's my view of what gnome-software is) advertising and enabling users to install something like a totally different desktop environment?
Rui
desktop mailing list desktop@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/desktop
On Fri, Jan 31, 2014 at 2:12 AM, Frank Murphy frankly3d@gmail.com wrote:
On Thu, 30 Jan 2014 16:08:10 -0500 Josh Boyer jwboyer@fedoraproject.org wrote:
I also don't think it excludes the possibility of different Workstation DEs in the future
I haven't read the final prd. Will initial default product allow for User Foo yum erase Default Desktop, yum install DE of choice. Whether by make make install or groupinstall DE
I don't think the WG has any plans to alter the low-level command line tools in Fedora, so what you describe should be possible. Actually removing the default DE would likely result in you running something the WG wouldn't consider Workstation, but it should be possible. If you want to use a DE other than the one Workstation picks for the basis of the product, you might be better off just using a Spin in the first place.
josh
On Fri, 31 Jan 2014 07:50:32 -0500 Josh Boyer jwboyer@fedoraproject.org wrote:
you might be better off just using a Spin in the first place.
Just allowing for worst case scenarios.
___ Regards, Frank www.frankly3d.com
----- Original Message -----
That is a good question and there is always the risk of 'overloading' the app with to many purposes. That said in the next version there will be support for installing fonts, codecs and ibus methods, so there is some precedence for installing things that is technically not an 'application'.
Hopefully some of the designers can chime in on this, I mean we could also considering putting this functionality somewhere else than the application installer if the designers thinks that is a better option.
I think creating a UI for this isn't one bit useful. It would probably be better served by separate spins (which joins up with the discussions currently on fedora-devel).
I mean, where would we stop for this? Allow changing the display manager from the UI? Making sure that those desktops can actually work with Wayland when GNOME and gdm switches to that?
It would severely limit the changes we can make to the workstation product, having it block on the non-primary desktops, a current problem we're trying to solve.
On Thu, 2014-01-30 at 12:55 -0800, Adam Williamson wrote:
Given the tension between the definition of a "Workstation Product" and the multiple desktop spins that I've identified on devel@ - i.e. that a "Workstation product" built around a single desktop occupies the 'desktop space', without accounting for alternative desktops - do you definitely want to go ahead with the model where the WS product is specifically associated with a single desktop and makes no attempt to somehow 'include' alternative desktops, or is it worth considering possible approaches that somehow account for alternatives? I realize it might be quite late to do that, but it seemed worth asking the question.
I don't think it makes any sense to build a product if you don't specify what it consists of. So yes, we need decide on a single set of packages that make up the Workstation. That is what you get if you install it. That is what the QA team will test to make sure it works. Etc.
That does not mean that there will be no 'accounting' for alternative desktops. I think Christian has maybe outlined before that there can be room for an option to install a different DE in addition, if it 'plays by the rules', which still have to be determined, but I would expect the minimum would be:
- appears as a session type on the login screen - uses common infrastructure where possible (things like localed, logind, etc) - does not interfere with the function of the core Workstation package set (ie the main DE)
It wouldn't be blocking, just like we wouldn't be blocking on any other piece of available software not working for a given release.
Christian
----- Original Message -----
From: "Bastien Nocera" bnocera@redhat.com To: "Discussions about development for the Fedora desktop" desktop@lists.fedoraproject.org Sent: Friday, January 31, 2014 2:25:53 PM Subject: Re: Underlying DE for the Workstation product
----- Original Message -----
That is a good question and there is always the risk of 'overloading' the app with to many purposes. That said in the next version there will be support for installing fonts, codecs and ibus methods, so there is some precedence for installing things that is technically not an 'application'.
Hopefully some of the designers can chime in on this, I mean we could also considering putting this functionality somewhere else than the application installer if the designers thinks that is a better option.
I think creating a UI for this isn't one bit useful. It would probably be better served by separate spins (which joins up with the discussions currently on fedora-devel).
I mean, where would we stop for this? Allow changing the display manager from the UI? Making sure that those desktops can actually work with Wayland when GNOME and gdm switches to that?
It would severely limit the changes we can make to the workstation product, having it block on the non-primary desktops, a current problem we're trying to solve. -- desktop mailing list desktop@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/desktop
Actually let me expand on my previous response.
First of all I sincerely hope we avoid 'spins' in the context they have been in Fedora up to now as I think that if we do continue with the the term Fedora will remain as undefinable as today, which is what I felt the fedora.next proposal actually wanted to get away from.
Secondly, as outlined in the PRD I do want there to be certain rules that any software available need to conform to in order to get made available in the UI. So I am not advocating a wild west scenario here.
Also since we are making a clear statement of a mandatory set of packages and DE, then if a given desktop environment is not ready for some reason for a given release then the user would get dropped back into the default upon upgrade.
We could also consider if there needs to be messaging about this in the installer, ie. this software is available, but we make no guarantees about it being available in the future.
Christian
----- Original Message -----
From: "Christian Schaller" cschalle@redhat.com To: "Discussions about development for the Fedora desktop" desktop@lists.fedoraproject.org Sent: Friday, January 31, 2014 2:35:30 PM Subject: Re: Underlying DE for the Workstation product
It wouldn't be blocking, just like we wouldn't be blocking on any other piece of available software not working for a given release.
Christian
----- Original Message -----
From: "Bastien Nocera" bnocera@redhat.com To: "Discussions about development for the Fedora desktop" desktop@lists.fedoraproject.org Sent: Friday, January 31, 2014 2:25:53 PM Subject: Re: Underlying DE for the Workstation product
----- Original Message -----
That is a good question and there is always the risk of 'overloading' the app with to many purposes. That said in the next version there will be support for installing fonts, codecs and ibus methods, so there is some precedence for installing things that is technically not an 'application'.
Hopefully some of the designers can chime in on this, I mean we could also considering putting this functionality somewhere else than the application installer if the designers thinks that is a better option.
I think creating a UI for this isn't one bit useful. It would probably be better served by separate spins (which joins up with the discussions currently on fedora-devel).
I mean, where would we stop for this? Allow changing the display manager from the UI? Making sure that those desktops can actually work with Wayland when GNOME and gdm switches to that?
It would severely limit the changes we can make to the workstation product, having it block on the non-primary desktops, a current problem we're trying to solve. -- desktop mailing list desktop@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/desktop
-- desktop mailing list desktop@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/desktop
On Fri, 2014-01-31 at 11:00 +0100, Rui Tiago Cação Matos wrote:
On 31 January 2014 10:47, Christian Schaller cschalle@redhat.com wrote:
Through the Software installer there will be other options available for install, like KDE. These will be installed alongside the default package set. We will not have a replace default option, just an add one, in order to ensure that the default package set can be targeted by 3rd parties.
This is an interesting one. Do we want an application installer (that's my view of what gnome-software is) advertising and enabling users to install something like a totally different desktop environment?
Aside from the question of correctness, I'm not sure it's going to be a hugely popular deployment method, and certainly not sufficient as the only one. A lot of people don't want to have to install GNOME in order to get to KDE.
So, it might be of use to some people to have a way of deploying other desktops from within GNOME - and they may benefit from "painting within the lines" of using the 'Workstation product', where they keep the 'mandatory packages', whatever they turn out to be, installed alongside their preferred DE - but I'm fairly sure it won't satisfy everyone, and we will still want to have some kind of a way for people to deploy a Fedora system with just KDE or just LXDE or just Cinnamon or just (whatever desktop they want).
On Fri, 2014-01-31 at 11:11 -0800, Adam Williamson wrote:
On Fri, 2014-01-31 at 11:00 +0100, Rui Tiago Cação Matos wrote:
On 31 January 2014 10:47, Christian Schaller cschalle@redhat.com wrote:
Through the Software installer there will be other options available for install, like KDE. These will be installed alongside the default package set. We will not have a replace default option, just an add one, in order to ensure that the default package set can be targeted by 3rd parties.
This is an interesting one. Do we want an application installer (that's my view of what gnome-software is) advertising and enabling users to install something like a totally different desktop environment?
Aside from the question of correctness, I'm not sure it's going to be a hugely popular deployment method, and certainly not sufficient as the only one. A lot of people don't want to have to install GNOME in order to get to KDE.
So, it might be of use to some people to have a way of deploying other desktops from within GNOME - and they may benefit from "painting within the lines" of using the 'Workstation product', where they keep the 'mandatory packages', whatever they turn out to be, installed alongside their preferred DE - but I'm fairly sure it won't satisfy everyone, and we will still want to have some kind of a way for people to deploy a Fedora system with just KDE or just LXDE or just Cinnamon or just (whatever desktop they want).
BTW, I think there's an interesting question emerging in the background here, and a rather useful consensus, in fact.
I think we can probably all agree on the following:
* The Workstation product should be free to define some criteria that you have to meet to be considered to be "running the Workstation product".
Right? Obviously that's what the WG wants, and I don't see a problem with that *in itself*. I don't think it would be fair to try and argue that Workstation WG must be required to consider people running KDE or Xfce or whatever as "running the Workstation product" if they don't *want* to. You folks seem to be discussing a mechanism/policy by which people can install extra desktops and still be considered to be "running the Workstation product", which is, I mean, A+ cool stuff!, but it's kind of an optional extra. *In theory* I think it would be legitimate for you to just say "To be considered to be running Fedora Workstation, you must be running GNOME" - again, I'm not suggesting that's *actually* what you want to do, I'm just illustrating a scenario.
So the interesting question becomes, what does it mean to *not* be "running the Workstation product"? (And, of course, by extension, the same applies to all the other products).
I don't know, maybe this isn't as interesting to anyone else, but to *me* this seems like a useful and productive way of conceptualizing things which actually makes me feel less worried about .next. We can define what expectations people can reasonably have if they're "running Fedora", what expectations people can reasonably have if they're "running Workstation", and so forth. And we certainly still have the opportunity to make the experience of just "running Fedora" be as good as it is today, and then making the Products spaces where you have interesting *extra* expectations.
If I look at it this way, things start looking kind of awesome. For instance, it *is* an interesting idea to say "if you meet the rules of "running Fedora Workstation", you can expect that third party software that complies with (these standards) will work". Looked at as an *extra* expectation that we provide via this "Product" space, that starts looking like a cool new thing.
So, yeah, I guess I'm feeling chirpier this morning. =) If we think of the Products as being spaces where we define certain requirements for the software/configuration set you're running, and *in exchange for that* we provide some extra expectations and functionalities that it just isn't plausible to provide in the "you can run any package set you like!" world - but we still explicitly say it's OK to just "run Fedora", with whatever package set you like, and still expect, you know, for yum to work and package updates to not conflict and all that good stuff...hey, that actually sounds like a pretty awesome world.
I can imagine we'll wind up having to make a *few* compromises, but I think we could make it work out really well.
So, am I right to look at it this way? Does that more or less line up with how the folks driving the Products see things? If so, expect less negative energy from me in future :P
* Warning: lengthy e-mail *
Hi,
I've read this thread quite carefully and hopefully haven't missed any mails to it. While quite a few people have come forward and proposed their DE of choice, I haven't really seen a mail which outlined what the criteria for choosing this DE is. Would it not be more helpful to decide on a set of criteria and then see what DE fits it best?
From the PRD[1], and the 4 use cases listed there, these are the characteristics that I think the chosen DE should possess. I fall into the university "student" + "independent developer" categories and I've mentioned my preferences with each criteria. All of this is based on my personal experience (and I am a happy GNOME3 user). Please feel free to add criteria/DEs/comments:
** Should be stable ** If we want students/devs to use the DE, they should be able to use it without it crashing again and again. We will of course QA our packages, but each upstream release should be stable in itself.
Students and developers do tend to have newer hardware. Such hardware should be able to support all our DEs. I don't think our chosen audience is one that only uses netbooks for example.
-> I think all the DEs Fedora provides currently are quite stable. I don't know about E yet, but I don't think that's a contender at this point.
** Active upstream, well maintained ** IMO, it's a bad idea to use a DE that doesn't have all it's sub components properly maintained. I'd prefer a DE, the upstream of which is active towards bugs and user feedback. This might not be relevant to the 4 use cases directly, but it is relevant to community members who'll look to work on this product and help improve it. It's also important in the long run that we work with an upstream that has plans and is moving forward to keep with tech advances.
-> Again, most DEs that Fedora provides have active upstreams. From what I read, I have the impression that MATE is sort of short on man power. (I could be wrong.) GNOME/KDE/XFCE/LXDE/Cinnamon seem to have active, organized communities that make regular, scheduled, timely releases. The desktops aren't just in maintenance mode, upstream continue to work on improving them. E had a release after *quite* a while IIRC.
** Well integrated application suite that includes: ** - mail client - chat client - calendar with google calendar etc support - task list - social networking clients - a photographic software that enables people to touch up their photographs and upload to popular image websites. - web browser (FF?) - printing support - backup/restore - music - games? - media-sharing/dlna?
These are necessities. These applications shouldn't just be available, they should go well with the rest of the environment.
Mail/chat/web/calendar clients are necessary for all our use cases. I tend to rely heavily on a task manager too (I don't consider Gnotes a task manager, I use gtg). University students and developers (more often than not), tend to make heavy use of social networking clients and camera software, even if just to give the photo a black/white transformation.
Backup/restore applications are good to have too, since more and more people have begun to use external hard drives to keep copies of sensitive data.
-> From my experience, GNOME/KDE are best in this regard, and I lean towards GNOME.
I think GOA is quite stable now, and setting up an account at one place and letting evolution/empathy/docs/photos use it is really convenient. I don't think there's a social networking client that uses GOA yet, but there are a few that use GTK/GTK3 (corebird is one that is under active development and quite stable now.) and look decent with the rest of the desktop. Since I maintain GTG and Hamster, both of which are working to integrate with EDS, I expect an even better experience in the short future, personally. (They can already communicate with each other using dbus). I also think the gnome search providers that provide results from: docs, contacts, applications, bijiben, wikipedia (extension), software (in 3.12 I think?), files, a browser (epiphany, not sure if this can be done with FF yet, but that would be *awesome*), even boxes; is pretty neat. Quite a few times, I can just type in the search box instead of opening nautilus at all.
KDE does have apps too: Kontact provides mail/contacts/tasks etc. last I checked. I don't know more details, unfortunately.
XFCE/MATE/LXDE/Cinnamon/E can all use apps from the repositories but they aren't as well integrated with the DE. For example, I would prefer to use either KDE or GNOME and their application suites rather than XFCE with Kontact installed on top of it.
I'm not considering Libreoffice here since I think we'd ship it irrespective of the DE.
** Easy to use configuration utilities ** - user setup - network - firewall - printer - input methods - accessibility?
We do have the system-config-* set of tools that would work with all DEs but they're not properly integrated. They're also not very actively maintained iirc with each desktop preferring to use their own tools. KDE/GNOME has them: date/time/printer/input-methods/network.
Do XFCE/MATE/LXDE/Cinnamon/E have dedicated configuration tools?
-> GNOME/KDE here too for me.
** Easy to use software manager ** We don't want nor expect all these users to use yum/dnf from the command line. For non IT related students, I think gnome-software is great. Apper/yumex are good too, but I like the idea of *applications* over *packages*, especially with people more used to such a UI owing to android/apple stores and such.
Deviating a bit here: It'll be awesome if we did have a similar webstore. We do have the packages app on Fedora infra, but again it's *packages* based. It's still quite useful though, since it works like a web frontend to `yum search`. But, you can't install a package quickly from the interface. It takes you to a package details page and then you'll probably have to go to Koji!? It's for the community and not users, apparently. https://apps.fedoraproject.org/packages/
Even the tagger app doesn't seem to have an "install" button.
I think gnome-software is to be extended to include reviewing and tagging capabilities, though?
-> I usually use dnf off the terminal, but I really think gnome-software is a win. Sure, people do disagree with some of its policies, like the "reboot to update" step, which I don't like too much myself, but people are not really averse to rebooting their systems for updates. (Windows set the tone. Mac doesn't always need a reboot I think, but it does on for some updates.) The behaviour could be tweaked but that's upstream's decision. It certainly isn't a blocker to me.
Apper/yumex aren't far behind so I wouldn't say I can't use them, but I do like the idea of gnome-software. Of course, we could always use gnome-software on top of KDE/LXDE/XFCE/Cinnamon/E. :) Will this pull in unwanted deps?
** Simple to use virtual client ** I saw this requirement in cases 3 and 4. I don't know if the WG has decided how to go about doing this yet.
-> I don't really have an app of choice here. virt-manager is great, but I think the current policy needs one to be an administrator to create a new vm. Gnome-boxes is quite nice and probably takes the cake from a simple to use point of view. I'm not aware of clients that integrate with the other DEs to a similar extent.
** Enterprise login ** I don't use this at all, so I'm not going to say anything about it. (Required in case 4.)
** Dev tools ** There are enough text editors that work well with each DE. Are there any dev tools specific to a certain DE? There are things like git viewers (gitg?) but they'll probably go well with all DEs.
-> No choice. I think dev tools will go well with most DEs.
** Hardware support ** - Multi monitor support - ?? I think most devs, either independent or in software companies tend to use at leasat 2 screens. Even students here at the robotics labs have 2 monitors. It'll be good to have. University students do end up presenting on a projector as part of courses at times too, so it'll be good to have that working.
-> No clue about the various DEs on this one. I've filed a couple of bugs against gnome-shell recently since I got my second monitor, though. Nothing major, just some RFEs which I thought were worth filing.
** Suspend/resume support ** Is this a DE issue, or a kernel/systemd issue?
** Easy to use UI ** There's been a debate on this one since the Gnome-shell design came out. A lot of people prefer the traditional desktop, customizable panel, window markers. Here, it's quite GNOME vs the rest. I feel this is a personal preference really. If we're looking at new users that are moving from Windows8 or users of Android/Apple phones, I expect them to get along quite well with GNOME. People who are used to their traditional systems will always find it difficult to start with GNOME.
-> I use a mostly pristine GNOME and it works for me. This is limited to me since I hardly use the mouse. (I navigate between all my windows using keyboard shortcuts). It's not easy to generalize.
** Should look good? (flashy?) ** This isn't a major requirement but anyone that's used mac/android/windows is sort of used to a shiny interface. University students do like to have flashy systems. Devs, not so much?
-> GNOME/KDE.
** ?? (please add/refine this criteria) ** Add more here!
Would it be easier to do this on a matrix on the wiki? It's not meant to be a popularity poll, but people who use different DEs could fill up how well their DEs fulfill the requirements. It's quite difficult for one user to know details about multiple DEs. Most of us stick to one that works for us.
For me, it's between GNOME and KDE. Since I'm well versed with GNOME and I really dig the integration I'd pick GNOME.
PS: If I've missed a DE, please do throw it into the mix.
[1] https://fedoraproject.org/wiki/Workstation/Workstation_PRD
Sending this out again. We seem to be discussing everything other than what the underlying DE should provide.
* Warning: lengthy e-mail *
Hi,
I've read this thread quite carefully and hopefully haven't missed any mails to it. While quite a few people have come forward and proposed their DE of choice, I haven't really seen a mail which outlined what the criteria for choosing this DE is. Would it not be more helpful to decide on a set of criteria and then see what DE fits it best?
From the PRD[1], and the 4 use cases listed there, these are the characteristics that I think the chosen DE should possess. I fall into the university "student" + "independent developer" categories and I've mentioned my preferences with each criteria. All of this is based on my personal experience (and I am a happy GNOME3 user). Please feel free to add criteria/DEs/comments:
** Should be stable ** If we want students/devs to use the DE, they should be able to use it without it crashing again and again. We will of course QA our packages, but each upstream release should be stable in itself.
Students and developers do tend to have newer hardware. Such hardware should be able to support all our DEs. I don't think our chosen audience is one that only uses netbooks for example.
-> I think all the DEs Fedora provides currently are quite stable. I don't know about E yet, but I don't think that's a contender at this point.
** Active upstream, well maintained ** IMO, it's a bad idea to use a DE that doesn't have all it's sub components properly maintained. I'd prefer a DE, the upstream of which is active towards bugs and user feedback. This might not be relevant to the 4 use cases directly, but it is relevant to community members who'll look to work on this product and help improve it. It's also important in the long run that we work with an upstream that has plans and is moving forward to keep with tech advances.
-> Again, most DEs that Fedora provides have active upstreams. From what I read, I have the impression that MATE is sort of short on man power. (I could be wrong.) GNOME/KDE/XFCE/LXDE/Cinnamon seem to have active, organized communities that make regular, scheduled, timely releases. The desktops aren't just in maintenance mode, upstream continue to work on improving them. E had a release after *quite* a while IIRC.
** Well integrated application suite that includes: ** - mail client - chat client - calendar with google calendar etc support - task list - social networking clients - a photographic software that enables people to touch up their photographs and upload to popular image websites. - web browser (FF?) - printing support - backup/restore - music - games? - media-sharing/dlna?
These are necessities. These applications shouldn't just be available, they should go well with the rest of the environment.
Mail/chat/web/calendar clients are necessary for all our use cases. I tend to rely heavily on a task manager too (I don't consider Gnotes a task manager, I use gtg). University students and developers (more often than not), tend to make heavy use of social networking clients and camera software, even if just to give the photo a black/white transformation.
Backup/restore applications are good to have too, since more and more people have begun to use external hard drives to keep copies of sensitive data.
-> From my experience, GNOME/KDE are best in this regard, and I lean towards GNOME.
I think GOA is quite stable now, and setting up an account at one place and letting evolution/empathy/docs/photos use it is really convenient. I don't think there's a social networking client that uses GOA yet, but there are a few that use GTK/GTK3 (corebird is one that is under active development and quite stable now.) and look decent with the rest of the desktop. Since I maintain GTG and Hamster, both of which are working to integrate with EDS, I expect an even better experience in the short future, personally. (They can already communicate with each other using dbus). I also think the gnome search providers that provide results from: docs, contacts, applications, bijiben, wikipedia (extension), software (in 3.12 I think?), files, a browser (epiphany, not sure if this can be done with FF yet, but that would be *awesome*), even boxes; is pretty neat. Quite a few times, I can just type in the search box instead of opening nautilus at all.
KDE does have apps too: Kontact provides mail/contacts/tasks etc. last I checked. I don't know more details, unfortunately.
XFCE/MATE/LXDE/Cinnamon/E can all use apps from the repositories but they aren't as well integrated with the DE. For example, I would prefer to use either KDE or GNOME and their application suites rather than XFCE with Kontact installed on top of it.
I'm not considering Libreoffice here since I think we'd ship it irrespective of the DE.
** Easy to use configuration utilities ** - user setup - network - firewall - printer - input methods - accessibility?
We do have the system-config-* set of tools that would work with all DEs but they're not properly integrated. They're also not very actively maintained iirc with each desktop preferring to use their own tools. KDE/GNOME has them: date/time/printer/input-methods/network.
Do XFCE/MATE/LXDE/Cinnamon/E have dedicated configuration tools?
-> GNOME/KDE here too for me.
** Easy to use software manager ** We don't want nor expect all these users to use yum/dnf from the command line. For non IT related students, I think gnome-software is great. Apper/yumex are good too, but I like the idea of *applications* over *packages*, especially with people more used to such a UI owing to android/apple stores and such.
Deviating a bit here: It'll be awesome if we did have a similar webstore. We do have the packages app on Fedora infra, but again it's *packages* based. It's still quite useful though, since it works like a web frontend to `yum search`. But, you can't install a package quickly from the interface. It takes you to a package details page and then you'll probably have to go to Koji!? It's for the community and not users, apparently. https://apps.fedoraproject.org/packages/
Even the tagger app doesn't seem to have an "install" button.
I think gnome-software is to be extended to include reviewing and tagging capabilities, though?
-> I usually use dnf off the terminal, but I really think gnome-software is a win. Sure, people do disagree with some of its policies, like the "reboot to update" step, which I don't like too much myself, but people are not really averse to rebooting their systems for updates. (Windows set the tone. Mac doesn't always need a reboot I think, but it does on for some updates.) The behaviour could be tweaked but that's upstream's decision. It certainly isn't a blocker to me.
Apper/yumex aren't far behind so I wouldn't say I can't use them, but I do like the idea of gnome-software. Of course, we could always use gnome-software on top of KDE/LXDE/XFCE/Cinnamon/E. :) Will this pull in unwanted deps?
** Simple to use virtual client ** I saw this requirement in cases 3 and 4. I don't know if the WG has decided how to go about doing this yet.
-> I don't really have an app of choice here. virt-manager is great, but I think the current policy needs one to be an administrator to create a new vm. Gnome-boxes is quite nice and probably takes the cake from a simple to use point of view. I'm not aware of clients that integrate with the other DEs to a similar extent.
** Enterprise login ** I don't use this at all, so I'm not going to say anything about it. (Required in case 4.)
** Dev tools ** There are enough text editors that work well with each DE. Are there any dev tools specific to a certain DE? There are things like git viewers (gitg?) but they'll probably go well with all DEs.
-> No choice. I think dev tools will go well with most DEs.
** Hardware support ** - Multi monitor support - ?? I think most devs, either independent or in software companies tend to use at leasat 2 screens. Even students here at the robotics labs have 2 monitors. It'll be good to have. University students do end up presenting on a projector as part of courses at times too, so it'll be good to have that working.
-> No clue about the various DEs on this one. I've filed a couple of bugs against gnome-shell recently since I got my second monitor, though. Nothing major, just some RFEs which I thought were worth filing.
** Suspend/resume support ** Is this a DE issue, or a kernel/systemd issue?
** Easy to use UI ** There's been a debate on this one since the Gnome-shell design came out. A lot of people prefer the traditional desktop, customizable panel, window markers. Here, it's quite GNOME vs the rest. I feel this is a personal preference really. If we're looking at new users that are moving from Windows8 or users of Android/Apple phones, I expect them to get along quite well with GNOME. People who are used to their traditional systems will always find it difficult to start with GNOME.
-> I use a mostly pristine GNOME and it works for me. This is limited to me since I hardly use the mouse. (I navigate between all my windows using keyboard shortcuts). It's not easy to generalize.
** Should look good? (flashy?) ** This isn't a major requirement but anyone that's used mac/android/windows is sort of used to a shiny interface. University students do like to have flashy systems. Devs, not so much?
-> GNOME/KDE.
** ?? (please add/refine this criteria) ** Add more here!
Would it be easier to do this on a matrix on the wiki? It's not meant to be a popularity poll, but people who use different DEs could fill up how well their DEs fulfill the requirements. It's quite difficult for one user to know details about multiple DEs. Most of us stick to one that works for us.
For me, it's between GNOME and KDE. Since I'm well versed with GNOME and I really dig the integration I'd pick GNOME.
PS: If I've missed a DE, please do throw it into the mix.
[1] https://fedoraproject.org/wiki/Workstation/Workstation_PRD
On Fri, 2014-01-31 at 04:47 -0500, Christian Schaller wrote:
[...]
I also took the liberty of adding a Technical Speification page to the wiki: https://fedoraproject.org/wiki/Workstation/Technical_Specification
It is currently more of a skeleton than anything else and I don't think I am the right person to write it, but I know Matthias and others are interested in helping flesh it out and add details, and hopefully we can have a good discussion on this list about potential items to include, wording and scoping etc.
Sorry for dragging my feet; I have now started to flesh this out.
desktop@lists.fedoraproject.org
-
Adam Williamson -
Ankur Sinha -
Bastien Nocera -
Christian Schaller -
Frank Murphy -
Josh Boyer -
Kalev Lember -
Lynn Dixon -
Matthias Clasen -
Rui Tiago Cação Matos