Hi All,
During the default DE discussions, a number of WG members expressed interest in keeping KDE as a release blocking DE for Workstation. QA is now asking FESCo about KDE's status as well in https://fedorahosted.org/fesco/ticket/1243
So if KDE is going to be a release blocking DE for Workstation, we need to figure out how exactly it gets installed and what manner it would be tested in. In the above ticket I came up with the following:
install the Workstation live image, install KDE through software-installer (if necessary), log into KDE from GDM after install, test
However, that was entirely off the top of my head. Would the live image be large enough to contain the KDE Workstation already or would a user/QA tester need to install it through the software-installer? What tests should be done? Etc.
Thoughts?
josh
On Tue, 2014-03-04 at 19:59 -0500, Josh Boyer wrote:
Hi All,
During the default DE discussions, a number of WG members expressed interest in keeping KDE as a release blocking DE for Workstation. QA is now asking FESCo about KDE's status as well in https://fedorahosted.org/fesco/ticket/1243
So if KDE is going to be a release blocking DE for Workstation, we need to figure out how exactly it gets installed and what manner it would be tested in. In the above ticket I came up with the following:
install the Workstation live image, install KDE through software-installer (if necessary), log into KDE from GDM after install, test
Yes, that is roughly what Christian outlined as the vision for alternative desktops and the workstation. Note that the software installer does not currently have the required functionality (install alternative desktops or other large sets of related software like, say, 'Ruby support').
I don't think that any of the alternative desktops should be release blocking, though.
However, that was entirely off the top of my head. Would the live image be large enough to contain the KDE Workstation already or would a user/QA tester need to install it through the software-installer? What tests should be done? Etc.
And I don't think there should be a 'KDE workstation'.
On Mar 4, 2014 8:35 PM, "Matthias Clasen" mclasen@redhat.com wrote:
On Tue, 2014-03-04 at 19:59 -0500, Josh Boyer wrote:
Hi All,
During the default DE discussions, a number of WG members expressed interest in keeping KDE as a release blocking DE for Workstation. QA is now asking FESCo about KDE's status as well in https://fedorahosted.org/fesco/ticket/1243
So if KDE is going to be a release blocking DE for Workstation, we need to figure out how exactly it gets installed and what manner it would be tested in. In the above ticket I came up with the following:
install the Workstation live image, install KDE through software-installer (if necessary), log into KDE from GDM after install, test
Yes, that is roughly what Christian outlined as the vision for alternative desktops and the workstation. Note that the software installer does not currently have the required functionality (install alternative desktops or other large sets of related software like, say, 'Ruby support').
So is that something feasible for the first release of Workstation or is that likely to come later?
I don't think that any of the alternative desktops should be release blocking, though.
I expect there will be differing opinions in the WG on that point.
However, that was entirely off the top of my head. Would the live image be large enough to contain the KDE Workstation already or would a user/QA tester need to install it through the software-installer? What tests should be done? Etc.
And I don't think there should be a 'KDE workstation'.
Sorry, I didn't mean to imply something separate. I meant "the commonly themed and integrated KDE DE".
josh
On Tue, 2014-03-04 at 19:59 -0500, Josh Boyer wrote:
Hi All,
During the default DE discussions, a number of WG members expressed interest in keeping KDE as a release blocking DE for Workstation. QA is now asking FESCo about KDE's status as well in https://fedorahosted.org/fesco/ticket/1243
So if KDE is going to be a release blocking DE for Workstation, we need to figure out how exactly it gets installed and what manner it would be tested in. In the above ticket I came up with the following:
install the Workstation live image, install KDE through software-installer (if necessary), log into KDE from GDM after install, test
However, that was entirely off the top of my head. Would the live image be large enough to contain the KDE Workstation already or would a user/QA tester need to install it through the software-installer? What tests should be done? Etc.
Thoughts?
This with my Personal Opinion hat on, not representing QA:
I'm not sure all/most people who actually want to use Fedora KDE are likely to be sold on doing it by downloading what they will see as 'GNOME', installing that, and then installing KDE on top of it. I think this will be fine for some folks, but there'll be a significant constituency which just wants a KDE image.
In fact we might be creating a bit of a problem, because I can see both "want KDE as an alternative desktop on top of the Workstation product" and "just want Fedora KDE" as two entirely legitimate and viable constituencies, which sort of means we've just created a bunch of extra work for ourselves. I'm not sure I see a clever magical solution to that, though. Engage brain cells...
On Tue, Mar 4, 2014 at 6:20 PM, Adam Williamson awilliam@redhat.com wrote:
This with my Personal Opinion hat on, not representing QA:
I'm not sure all/most people who actually want to use Fedora KDE are likely to be sold on doing it by downloading what they will see as 'GNOME', installing that, and then installing KDE on top of it. I think this will be fine for some folks, but there'll be a significant constituency which just wants a KDE image.
In fact we might be creating a bit of a problem, because I can see both "want KDE as an alternative desktop on top of the Workstation product" and "just want Fedora KDE" as two entirely legitimate and viable constituencies, which sort of means we've just created a bunch of extra work for ourselves. I'm not sure I see a clever magical solution to that, though. Engage brain cells...
+1
But I don't really know of a way to add to this discussion besides pointing you to spins.
Dan
On 5 March 2014 02:13, Josh Boyer jwboyer@fedoraproject.org wrote:
So is that something feasible for the first release of Workstation or is that likely to come later?
It's quite an easy thing to do, I just need mockups on how it should look.
Richard.
----- Original Message -----
From: "Adam Williamson" awilliam@redhat.com To: "Discussions about development for the Fedora desktop" desktop@lists.fedoraproject.org Sent: Wednesday, March 5, 2014 3:20:06 AM Subject: Re: KDE integration/status for Workstation
On Tue, 2014-03-04 at 19:59 -0500, Josh Boyer wrote:
Hi All,
During the default DE discussions, a number of WG members expressed interest in keeping KDE as a release blocking DE for Workstation. QA is now asking FESCo about KDE's status as well in https://fedorahosted.org/fesco/ticket/1243
So if KDE is going to be a release blocking DE for Workstation, we need to figure out how exactly it gets installed and what manner it would be tested in. In the above ticket I came up with the following:
install the Workstation live image, install KDE through software-installer (if necessary), log into KDE from GDM after install, test
However, that was entirely off the top of my head. Would the live image be large enough to contain the KDE Workstation already or would a user/QA tester need to install it through the software-installer? What tests should be done? Etc.
Thoughts?
This with my Personal Opinion hat on, not representing QA:
I'm not sure all/most people who actually want to use Fedora KDE are likely to be sold on doing it by downloading what they will see as 'GNOME', installing that, and then installing KDE on top of it. I think this will be fine for some folks, but there'll be a significant constituency which just wants a KDE image.
In fact we might be creating a bit of a problem, because I can see both "want KDE as an alternative desktop on top of the Workstation product" and "just want Fedora KDE" as two entirely legitimate and viable constituencies, which sort of means we've just created a bunch of extra work for ourselves. I'm not sure I see a clever magical solution to that, though. Engage brain cells... --
Well my take is that there will never be a solution that makes everyone happy, but in my mind the expectation here has always been that the people who would not be happy about the proposed solution would end up focusing their energy on doing a remix. Because there isn't really a way we can make a product and at the same time be a solution for people who essentially want a different product.
Christian
----- Original Message -----
----- Original Message -----
From: "Adam Williamson" awilliam@redhat.com To: "Discussions about development for the Fedora desktop" desktop@lists.fedoraproject.org Sent: Wednesday, March 5, 2014 3:20:06 AM Subject: Re: KDE integration/status for Workstation
On Tue, 2014-03-04 at 19:59 -0500, Josh Boyer wrote:
Hi All,
During the default DE discussions, a number of WG members expressed interest in keeping KDE as a release blocking DE for Workstation. QA is now asking FESCo about KDE's status as well in https://fedorahosted.org/fesco/ticket/1243
So if KDE is going to be a release blocking DE for Workstation, we need to figure out how exactly it gets installed and what manner it would be tested in. In the above ticket I came up with the following:
install the Workstation live image, install KDE through software-installer (if necessary), log into KDE from GDM after install, test
However, that was entirely off the top of my head. Would the live image be large enough to contain the KDE Workstation already or would a user/QA tester need to install it through the software-installer? What tests should be done? Etc.
Thoughts?
This with my Personal Opinion hat on, not representing QA:
I'm not sure all/most people who actually want to use Fedora KDE are likely to be sold on doing it by downloading what they will see as 'GNOME', installing that, and then installing KDE on top of it. I think this will be fine for some folks, but there'll be a significant constituency which just wants a KDE image.
In fact we might be creating a bit of a problem, because I can see both "want KDE as an alternative desktop on top of the Workstation product" and "just want Fedora KDE" as two entirely legitimate and viable constituencies, which sort of means we've just created a bunch of extra work for ourselves. I'm not sure I see a clever magical solution to that, though. Engage brain cells... --
Well my take is that there will never be a solution that makes everyone happy, but in my mind the expectation here has always been that the people who would not be happy about the proposed solution would end up focusing their energy on doing a remix. Because there isn't really a way we can make a product and at the same time be a solution for people who essentially want a different product.
Yep, and that probably leads back to the question if we want or don't want to provide more products than we currently have so far (three). And to be honest, I like idea having more products but with very high bar (that should apply for all products) - to avoid a situation of having dozens of products but still be inclusive.
There was FESCo ticket https://fedorahosted.org/fesco/ticket/1197 - it's now closed with resolution: "AGREED: Close for now until the current products are released at least once and/or anyone proposes a new product" and it's states that the bar should be relatively high too.
KDE would be a great example of such first product - pretty strong KDE SIG team, that consists of a few very active community members and Red Hatters, proved to be able to solve release blockers (and pretty tough release blockers), with quite a big user base and last few years known as one of the best and most polished KDE experience (from distribution POV).
Steps 1. establish formal KDE WG from the initial KDE SIG members 2. create formal PRD 3. create Technical Specifications (based on WS Tech Specs or future Base Tech Specs) 4. show commitment, provide resources for other teams - QA (test cases, test days, help with release matrix coverage), marketing etc. 5. get approved - does not have to be for F21, start line would be pretty late but do not block it, if KDE WG would be able to make it 6. profit!
For the idea of having for example Scientific products based on KDE - I don't like it as these topic specific spin like things should be desktop agnostic and really done on top of existing desktops in Software Installer (GNOME, KDE whatever). And create spins/formulas whatever only when it's really needed (but then it would be more remix).
One another thought - it could serve as a place for desktops that would not fulfil WS WG requirements but could fulfil KDE WG requirements (there are a lot of new Qt/KDE based desktops appearing recently)...
CC'ing KDE mailing list. Due to some other stuff, I'm not very active recently in discussion but initially KDE SIG was more inclined to own product, then I saw a few people more inclined being part of WS WG and now I think it's clearly steering back to spin/product. Remix would be suicide - even KDE SIG is pretty big team, active, with healthy community, it would mean a lot of duplication etc.
Jaroslav
Christian
-- desktop mailing list desktop@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/desktop
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 03/04/2014 09:20 PM, Adam Williamson wrote:
This with my Personal Opinion hat on, not representing QA:
I'm not sure all/most people who actually want to use Fedora KDE are likely to be sold on doing it by downloading what they will see as 'GNOME', installing that, and then installing KDE on top of it. I think this will be fine for some folks, but there'll be a significant constituency which just wants a KDE image.
In fact we might be creating a bit of a problem, because I can see both "want KDE as an alternative desktop on top of the Workstation product" and "just want Fedora KDE" as two entirely legitimate and viable constituencies, which sort of means we've just created a bunch of extra work for ourselves. I'm not sure I see a clever magical solution to that, though. Engage brain cells...
I'd suggest that for the Fedora Workstation, we declare that KDE is release-blocking *as an optional component atop the Workstation*.
For the constituency that wants a pristine KDE environment, they may choose to maintain their own spin, but *that spin* should be removed from the Fedora QA's matrix. (Fedora QA should focus on the Workstation add-on). The KDE SIG likely has its own sufficient resources for testing their pure-KDE spin.
On Wed, Mar 5, 2014 at 8:48 AM, Stephen Gallagher sgallagh@redhat.com wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 03/04/2014 09:20 PM, Adam Williamson wrote:
This with my Personal Opinion hat on, not representing QA:
I'm not sure all/most people who actually want to use Fedora KDE are likely to be sold on doing it by downloading what they will see as 'GNOME', installing that, and then installing KDE on top of it. I think this will be fine for some folks, but there'll be a significant constituency which just wants a KDE image.
In fact we might be creating a bit of a problem, because I can see both "want KDE as an alternative desktop on top of the Workstation product" and "just want Fedora KDE" as two entirely legitimate and viable constituencies, which sort of means we've just created a bunch of extra work for ourselves. I'm not sure I see a clever magical solution to that, though. Engage brain cells...
I'd suggest that for the Fedora Workstation, we declare that KDE is release-blocking *as an optional component atop the Workstation*.
Please explain this further. Having an optional component be release blocking is making my head hurt.
josh
----- Original Message -----
From: "Josh Boyer" jwboyer@fedoraproject.org To: "Discussions about development for the Fedora desktop" desktop@lists.fedoraproject.org Sent: Wednesday, March 5, 2014 3:20:44 PM Subject: Re: KDE integration/status for Workstation
On Wed, Mar 5, 2014 at 8:48 AM, Stephen Gallagher sgallagh@redhat.com wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 03/04/2014 09:20 PM, Adam Williamson wrote:
This with my Personal Opinion hat on, not representing QA:
I'm not sure all/most people who actually want to use Fedora KDE are likely to be sold on doing it by downloading what they will see as 'GNOME', installing that, and then installing KDE on top of it. I think this will be fine for some folks, but there'll be a significant constituency which just wants a KDE image.
In fact we might be creating a bit of a problem, because I can see both "want KDE as an alternative desktop on top of the Workstation product" and "just want Fedora KDE" as two entirely legitimate and viable constituencies, which sort of means we've just created a bunch of extra work for ourselves. I'm not sure I see a clever magical solution to that, though. Engage brain cells...
I'd suggest that for the Fedora Workstation, we declare that KDE is release-blocking *as an optional component atop the Workstation*.
Please explain this further. Having an optional component be release blocking is making my head hurt.
josh
At the risk of misrepresenting Stephen I think what he means, and what I agree with is that we declare that KDE is blocking as in 'it should work' before we do a given release, but it is optional to install for the end user.
Christian
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 03/05/2014 09:20 AM, Josh Boyer wrote:
On Wed, Mar 5, 2014 at 8:48 AM, Stephen Gallagher sgallagh@redhat.com wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 03/04/2014 09:20 PM, Adam Williamson wrote:
This with my Personal Opinion hat on, not representing QA:
I'm not sure all/most people who actually want to use Fedora KDE are likely to be sold on doing it by downloading what they will see as 'GNOME', installing that, and then installing KDE on top of it. I think this will be fine for some folks, but there'll be a significant constituency which just wants a KDE image.
In fact we might be creating a bit of a problem, because I can see both "want KDE as an alternative desktop on top of the Workstation product" and "just want Fedora KDE" as two entirely legitimate and viable constituencies, which sort of means we've just created a bunch of extra work for ourselves. I'm not sure I see a clever magical solution to that, though. Engage brain cells...
I'd suggest that for the Fedora Workstation, we declare that KDE is release-blocking *as an optional component atop the Workstation*.
Please explain this further. Having an optional component be release blocking is making my head hurt.
What I mean is that if I boot the live media, install from it and select "KDE" during that install process, then having a usable system after the post-install reboot is blocking.
Is that more clear?
On Wed, Mar 5, 2014 at 9:24 AM, Stephen Gallagher sgallagh@redhat.com wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 03/05/2014 09:20 AM, Josh Boyer wrote:
On Wed, Mar 5, 2014 at 8:48 AM, Stephen Gallagher sgallagh@redhat.com wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 03/04/2014 09:20 PM, Adam Williamson wrote:
This with my Personal Opinion hat on, not representing QA:
I'm not sure all/most people who actually want to use Fedora KDE are likely to be sold on doing it by downloading what they will see as 'GNOME', installing that, and then installing KDE on top of it. I think this will be fine for some folks, but there'll be a significant constituency which just wants a KDE image.
In fact we might be creating a bit of a problem, because I can see both "want KDE as an alternative desktop on top of the Workstation product" and "just want Fedora KDE" as two entirely legitimate and viable constituencies, which sort of means we've just created a bunch of extra work for ourselves. I'm not sure I see a clever magical solution to that, though. Engage brain cells...
I'd suggest that for the Fedora Workstation, we declare that KDE is release-blocking *as an optional component atop the Workstation*.
Please explain this further. Having an optional component be release blocking is making my head hurt.
What I mean is that if I boot the live media, install from it and select "KDE" during that install process, then having a usable system after the post-install reboot is blocking.
Is that more clear?
Er, yes and no. I think I see what you're getting at though and that's fine.
josh
----- Original Message -----
From: "Stephen Gallagher" sgallagh@redhat.com To: "Discussions about development for the Fedora desktop" desktop@lists.fedoraproject.org Sent: Wednesday, March 5, 2014 3:24:29 PM Subject: Re: KDE integration/status for Workstation
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 03/05/2014 09:20 AM, Josh Boyer wrote:
On Wed, Mar 5, 2014 at 8:48 AM, Stephen Gallagher sgallagh@redhat.com wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 03/04/2014 09:20 PM, Adam Williamson wrote:
This with my Personal Opinion hat on, not representing QA:
I'm not sure all/most people who actually want to use Fedora KDE are likely to be sold on doing it by downloading what they will see as 'GNOME', installing that, and then installing KDE on top of it. I think this will be fine for some folks, but there'll be a significant constituency which just wants a KDE image.
In fact we might be creating a bit of a problem, because I can see both "want KDE as an alternative desktop on top of the Workstation product" and "just want Fedora KDE" as two entirely legitimate and viable constituencies, which sort of means we've just created a bunch of extra work for ourselves. I'm not sure I see a clever magical solution to that, though. Engage brain cells...
I'd suggest that for the Fedora Workstation, we declare that KDE is release-blocking *as an optional component atop the Workstation*.
Please explain this further. Having an optional component be release blocking is making my head hurt.
What I mean is that if I boot the live media, install from it and select "KDE" during that install process, then having a usable system after the post-install reboot is blocking.
Is that more clear?
Well there would be no such option though. There is no package selection planned for the workstation install. All optional packages are meant to be installed through the Software installer.
Christian
On Wed, Mar 5, 2014 at 9:22 AM, Christian Schaller cschalle@redhat.com wrote:
----- Original Message -----
From: "Josh Boyer" jwboyer@fedoraproject.org To: "Discussions about development for the Fedora desktop" desktop@lists.fedoraproject.org Sent: Wednesday, March 5, 2014 3:20:44 PM Subject: Re: KDE integration/status for Workstation
On Wed, Mar 5, 2014 at 8:48 AM, Stephen Gallagher sgallagh@redhat.com wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 03/04/2014 09:20 PM, Adam Williamson wrote:
This with my Personal Opinion hat on, not representing QA:
I'm not sure all/most people who actually want to use Fedora KDE are likely to be sold on doing it by downloading what they will see as 'GNOME', installing that, and then installing KDE on top of it. I think this will be fine for some folks, but there'll be a significant constituency which just wants a KDE image.
In fact we might be creating a bit of a problem, because I can see both "want KDE as an alternative desktop on top of the Workstation product" and "just want Fedora KDE" as two entirely legitimate and viable constituencies, which sort of means we've just created a bunch of extra work for ourselves. I'm not sure I see a clever magical solution to that, though. Engage brain cells...
I'd suggest that for the Fedora Workstation, we declare that KDE is release-blocking *as an optional component atop the Workstation*.
Please explain this further. Having an optional component be release blocking is making my head hurt.
josh
At the risk of misrepresenting Stephen I think what he means, and what I agree with is that we declare that KDE is blocking as in 'it should work' before we do a given release, but it is optional to install for the end user.
OK, right. That's essentially what I was suggesting with the testing steps I initially threw out there. At least Matthias seems to think we shouldn't block in this manner, so it seems we have some discussion to do within the WG.
josh
On Wed, Mar 5, 2014 at 7:16 AM, Jaroslav Reznik jreznik@redhat.com wrote:
----- Original Message -----
----- Original Message -----
From: "Adam Williamson" awilliam@redhat.com To: "Discussions about development for the Fedora desktop" desktop@lists.fedoraproject.org Sent: Wednesday, March 5, 2014 3:20:06 AM Subject: Re: KDE integration/status for Workstation
On Tue, 2014-03-04 at 19:59 -0500, Josh Boyer wrote:
Hi All,
During the default DE discussions, a number of WG members expressed interest in keeping KDE as a release blocking DE for Workstation. QA is now asking FESCo about KDE's status as well in https://fedorahosted.org/fesco/ticket/1243
So if KDE is going to be a release blocking DE for Workstation, we need to figure out how exactly it gets installed and what manner it would be tested in. In the above ticket I came up with the following:
install the Workstation live image, install KDE through software-installer (if necessary), log into KDE from GDM after install, test
However, that was entirely off the top of my head. Would the live image be large enough to contain the KDE Workstation already or would a user/QA tester need to install it through the software-installer? What tests should be done? Etc.
Thoughts?
This with my Personal Opinion hat on, not representing QA:
I'm not sure all/most people who actually want to use Fedora KDE are likely to be sold on doing it by downloading what they will see as 'GNOME', installing that, and then installing KDE on top of it. I think this will be fine for some folks, but there'll be a significant constituency which just wants a KDE image.
In fact we might be creating a bit of a problem, because I can see both "want KDE as an alternative desktop on top of the Workstation product" and "just want Fedora KDE" as two entirely legitimate and viable constituencies, which sort of means we've just created a bunch of extra work for ourselves. I'm not sure I see a clever magical solution to that, though. Engage brain cells... --
Well my take is that there will never be a solution that makes everyone happy, but in my mind the expectation here has always been that the people who would not be happy about the proposed solution would end up focusing their energy on doing a remix. Because there isn't really a way we can make a product and at the same time be a solution for people who essentially want a different product.
Yep, and that probably leads back to the question if we want or don't want to provide more products than we currently have so far (three). And to be honest, I like idea having more products but with very high bar (that should apply for all products) - to avoid a situation of having dozens of products but still be inclusive.
All of your subsequent email is perfectly fine, but this thread is focused on what we want to do with KDE in the Workstation product. If you want to advocate for a KDE product, please take that discussion to the devel list. I don't want to derail this conversation, and this list isn't the right place to discuss what you're suggesting.
CC'ing KDE mailing list. Due to some other stuff, I'm not very active recently in discussion but initially KDE SIG was more inclined to own product, then I saw a few people more inclined being part of WS WG and now I think it's clearly steering back to spin/product. Remix would be suicide - even KDE SIG is pretty big team, active, with healthy community, it would mean a lot of duplication etc.
I think Christian actually meant spin, not remix, in the Fedora definition of those terms.
josh
----- Original Message -----
On Wed, Mar 5, 2014 at 7:16 AM, Jaroslav Reznik jreznik@redhat.com wrote:
----- Original Message -----
----- Original Message -----
From: "Adam Williamson" awilliam@redhat.com To: "Discussions about development for the Fedora desktop" desktop@lists.fedoraproject.org Sent: Wednesday, March 5, 2014 3:20:06 AM Subject: Re: KDE integration/status for Workstation
On Tue, 2014-03-04 at 19:59 -0500, Josh Boyer wrote:
Hi All,
During the default DE discussions, a number of WG members expressed interest in keeping KDE as a release blocking DE for Workstation. QA is now asking FESCo about KDE's status as well in https://fedorahosted.org/fesco/ticket/1243
So if KDE is going to be a release blocking DE for Workstation, we need to figure out how exactly it gets installed and what manner it would be tested in. In the above ticket I came up with the following:
install the Workstation live image, install KDE through software-installer (if necessary), log into KDE from GDM after install, test
However, that was entirely off the top of my head. Would the live image be large enough to contain the KDE Workstation already or would a user/QA tester need to install it through the software-installer? What tests should be done? Etc.
Thoughts?
This with my Personal Opinion hat on, not representing QA:
I'm not sure all/most people who actually want to use Fedora KDE are likely to be sold on doing it by downloading what they will see as 'GNOME', installing that, and then installing KDE on top of it. I think this will be fine for some folks, but there'll be a significant constituency which just wants a KDE image.
In fact we might be creating a bit of a problem, because I can see both "want KDE as an alternative desktop on top of the Workstation product" and "just want Fedora KDE" as two entirely legitimate and viable constituencies, which sort of means we've just created a bunch of extra work for ourselves. I'm not sure I see a clever magical solution to that, though. Engage brain cells... --
Well my take is that there will never be a solution that makes everyone happy, but in my mind the expectation here has always been that the people who would not be happy about the proposed solution would end up focusing their energy on doing a remix. Because there isn't really a way we can make a product and at the same time be a solution for people who essentially want a different product.
Yep, and that probably leads back to the question if we want or don't want to provide more products than we currently have so far (three). And to be honest, I like idea having more products but with very high bar (that should apply for all products) - to avoid a situation of having dozens of products but still be inclusive.
All of your subsequent email is perfectly fine, but this thread is focused on what we want to do with KDE in the Workstation product. If you want to advocate for a KDE product, please take that discussion to the devel list. I don't want to derail this conversation, and this list isn't the right place to discuss what you're suggesting.
Well, I really think the answer to the question , if KDE - as vague as KDE stands here - is release blocking (and now doesn't matter if as part of WS product, spin or even separate product) depends on the way, how it's going to be delivered.
Before we know it, it's non sense to discuss if having an option in the Software Installer is blocking thing and how to treat it. And I'd still say close coordination between WS WG and KDE SIG is the must.
On the other hand, I'd not object going even more forward and have WS product as a GNOME desktop but with desktop agnostic set of applications offered in default installation. I think that's the way Christian thinks about it - better integration of both of these worlds (from theming, to accessibility). Once this part is done, workstation product could be open to serve as real one, not one specific desktop showcase (and doesn't matter which one, I really don't care in such case) product. Then, even that post install does not have to be release blocking and only optional thing for people who really want the Plasma desktop. From what I understand from some WS WG folks, this could be preferred option but I heard, there were objections to have kdelibs in the tech specs. With move to KDE Frameworks 5, it could be raised again, as it's going to be more Qt 5 add on modules than old fashioned monolithic kdelibs.
Jaroslav
----- Original Message -----
From: "Josh Boyer" jwboyer@fedoraproject.org To: "Discussions about development for the Fedora desktop" desktop@lists.fedoraproject.org Cc: "KDE on Fedora discussion" kde@lists.fedoraproject.org Sent: Wednesday, March 5, 2014 3:29:42 PM Subject: Re: KDE integration/status for Workstation
On Wed, Mar 5, 2014 at 7:16 AM, Jaroslav Reznik jreznik@redhat.com wrote:
----- Original Message -----
----- Original Message -----
From: "Adam Williamson" awilliam@redhat.com To: "Discussions about development for the Fedora desktop" desktop@lists.fedoraproject.org Sent: Wednesday, March 5, 2014 3:20:06 AM Subject: Re: KDE integration/status for Workstation
On Tue, 2014-03-04 at 19:59 -0500, Josh Boyer wrote:
Hi All,
During the default DE discussions, a number of WG members expressed interest in keeping KDE as a release blocking DE for Workstation. QA is now asking FESCo about KDE's status as well in https://fedorahosted.org/fesco/ticket/1243
So if KDE is going to be a release blocking DE for Workstation, we need to figure out how exactly it gets installed and what manner it would be tested in. In the above ticket I came up with the following:
install the Workstation live image, install KDE through software-installer (if necessary), log into KDE from GDM after install, test
However, that was entirely off the top of my head. Would the live image be large enough to contain the KDE Workstation already or would a user/QA tester need to install it through the software-installer? What tests should be done? Etc.
Thoughts?
This with my Personal Opinion hat on, not representing QA:
I'm not sure all/most people who actually want to use Fedora KDE are likely to be sold on doing it by downloading what they will see as 'GNOME', installing that, and then installing KDE on top of it. I think this will be fine for some folks, but there'll be a significant constituency which just wants a KDE image.
In fact we might be creating a bit of a problem, because I can see both "want KDE as an alternative desktop on top of the Workstation product" and "just want Fedora KDE" as two entirely legitimate and viable constituencies, which sort of means we've just created a bunch of extra work for ourselves. I'm not sure I see a clever magical solution to that, though. Engage brain cells... --
Well my take is that there will never be a solution that makes everyone happy, but in my mind the expectation here has always been that the people who would not be happy about the proposed solution would end up focusing their energy on doing a remix. Because there isn't really a way we can make a product and at the same time be a solution for people who essentially want a different product.
Yep, and that probably leads back to the question if we want or don't want to provide more products than we currently have so far (three). And to be honest, I like idea having more products but with very high bar (that should apply for all products) - to avoid a situation of having dozens of products but still be inclusive.
All of your subsequent email is perfectly fine, but this thread is focused on what we want to do with KDE in the Workstation product. If you want to advocate for a KDE product, please take that discussion to the devel list. I don't want to derail this conversation, and this list isn't the right place to discuss what you're suggesting.
CC'ing KDE mailing list. Due to some other stuff, I'm not very active recently in discussion but initially KDE SIG was more inclined to own product, then I saw a few people more inclined being part of WS WG and now I think it's clearly steering back to spin/product. Remix would be suicide - even KDE SIG is pretty big team, active, with healthy community, it would mean a lot of duplication etc.
I think Christian actually meant spin, not remix, in the Fedora definition of those terms.
I actually did mean remix, but it of course do boil down to what spin vs remix actually ends up meaning. I do realize that they are both somewhat loosely defined terms with different people putting different meaning to them.
So in general I am strongly against a proliferation of products in the sense that I see the general Fedora download experience to be a page presenting you with 3 easily identifiable and zero overlap options (desktop, server, clould). One could argue that one could add a few more options there for other clearly distinct usecases, but in general if the page turns into a long list of options, we have lost the point doing the products to begin with.
Remixes for me is more along the lines of what is happening in the Ubuntu world, where all remixes/spins or whatever you want to call them are branded separately, making the distinction clear of what is the core product and what is derivatives. (That said I would suggest we do not do the alternative branding along the same lines as Ubuntu did making them all variations of the Fedora name.)
Anyway, I guess this discussion belongs on -devel and not here.
Christian
On Wed, 2014-03-05 at 08:48 -0500, Stephen Gallagher wrote:
On 03/04/2014 09:20 PM, Adam Williamson wrote:
This with my Personal Opinion hat on, not representing QA:
I'm not sure all/most people who actually want to use Fedora KDE are likely to be sold on doing it by downloading what they will see as 'GNOME', installing that, and then installing KDE on top of it. I think this will be fine for some folks, but there'll be a significant constituency which just wants a KDE image.
In fact we might be creating a bit of a problem, because I can see both "want KDE as an alternative desktop on top of the Workstation product" and "just want Fedora KDE" as two entirely legitimate and viable constituencies, which sort of means we've just created a bunch of extra work for ourselves. I'm not sure I see a clever magical solution to that, though. Engage brain cells...
I'd suggest that for the Fedora Workstation, we declare that KDE is release-blocking *as an optional component atop the Workstation*.
For the constituency that wants a pristine KDE environment, they may choose to maintain their own spin, but *that spin* should be removed from the Fedora QA's matrix. (Fedora QA should focus on the Workstation add-on). The KDE SIG likely has its own sufficient resources for testing their pure-KDE spin.
Well, as we do things now, QA and the KDE SIG collaborate on testing KDE under the current model (you can get it in two ways - live image or DVD - but they're very similar and don't need to be tested entirely separately). This approach still effectively nearly doubles the workload, because it envisages QA doing a full round of testing on KDE-atop-Workstation, and KDE SIG doing a full round of testing on KDE-spin.
On Wed, 2014-03-05 at 09:29 -0500, Josh Boyer wrote:
Yep, and that probably leads back to the question if we want or don't want to provide more products than we currently have so far (three). And to be honest, I like idea having more products but with very high bar (that should apply for all products) - to avoid a situation of having dozens of products but still be inclusive.
All of your subsequent email is perfectly fine, but this thread is focused on what we want to do with KDE in the Workstation product. If you want to advocate for a KDE product, please take that discussion to the devel list. I don't want to derail this conversation, and this list isn't the right place to discuss what you're suggesting.
Well. This thread was sparked partly by my FESCo ticket, which is written from a project-wide perspective, not a product/WG-specific one. The project as a whole has skin in this game: it is important to the project how the project chooses to deliver KDE, and if the answer is 'in more than one way', the relative importance of those ways. The Workstation WG is free to integrate KDE into their product in any appropriate way, of course, but I think the project *as a whole* has a say in whether that's our primary method of distributing KDE - how heavily it's promoted, how hard it's tested, and whether it blocks our releases. In that context, the existence or otherwise and relative importance of a KDE spin is obviously significant.
On Wed, 2014-03-05 at 09:46 -0500, Christian Schaller wrote:
I think Christian actually meant spin, not remix, in the Fedora definition of those terms.
I actually did mean remix, but it of course do boil down to what spin vs remix actually ends up meaning. I do realize that they are both somewhat loosely defined terms with different people putting different meaning to them.
Eh? No, they aren't. They're quite clearly defined.
https://spins.fedoraproject.org/about https://fedoraproject.org/wiki/Remix
a 'spin' is a much stricter category than a remix: basically it gets to choose its package set *from within the official repositories* and do certain types of customization you can do via a kickstart file, and that's it. Spins are pretty tied in to Fedora releases, pretty 'official' things.
a 'remix' is much looser: "a combination of Fedora software, with or without add-ons, that can be created by any community member at any time to produce interesting and compelling products." Anyone can build one, it doesn't have to be a kickstart-based live image, and it can include whatever third party software you like.
The Remix page has a concise summary:
"Are Remixes and Spins different?
Yes. A Remix can be created by anyone with our tools, and labeled with the special "Fedora Remix" mark as set out in that logo's guidelines page. A Spin, on the other hand, contains only 100% Fedora software, not combined with any other third-party software, has been approved by the Fedora Spins_SIG, and granted the "Fedora" name by the Fedora Project Board. Spins are usually carried on our BitTorrent server and official mirrors."
On Wed, Mar 5, 2014 at 12:58 PM, Adam Williamson awilliam@redhat.com wrote:
On Wed, 2014-03-05 at 09:29 -0500, Josh Boyer wrote:
Yep, and that probably leads back to the question if we want or don't want to provide more products than we currently have so far (three). And to be honest, I like idea having more products but with very high bar (that should apply for all products) - to avoid a situation of having dozens of products but still be inclusive.
All of your subsequent email is perfectly fine, but this thread is focused on what we want to do with KDE in the Workstation product. If you want to advocate for a KDE product, please take that discussion to the devel list. I don't want to derail this conversation, and this list isn't the right place to discuss what you're suggesting.
Well. This thread was sparked partly by my FESCo ticket, which is written from a project-wide perspective, not a product/WG-specific one.
Yes.
The project as a whole has skin in this game: it is important to the project how the project chooses to deliver KDE, and if the answer is 'in more than one way', the relative importance of those ways. The
Yes.
Workstation WG is free to integrate KDE into their product in any appropriate way, of course, but I think the project *as a whole* has a say in whether that's our primary method of distributing KDE - how heavily it's promoted, how hard it's tested, and whether it blocks our releases. In that context, the existence or otherwise and relative importance of a KDE spin is obviously significant.
Yes.
However, THIS LIST and THIS THREAD are not the appropriate places to be having that discussion because you're missing EVERYONE ELSE that actually has skin in the game. As I said, this is the thread for "what does Workstation want to do with KDE". The broader discussion needs to be hashed out on the devel list and/or the FESCo ticket.
Put another way, Workstation has said for a while now it wants to include KDE. We need to figure out as a product how we're going to do that and what we need to do to accomplish it. This is very much supposed to be the "Workstation WG is free to integrate KDE into their product in any appropriate way..." part.
josh
Jaroslav Reznik wrote:
I heard, there were objections to have kdelibs in the tech specs. With move to KDE Frameworks 5, it could be raised again, as it's going to be more Qt 5 add on modules than old fashioned monolithic kdelibs.
(Sorry for chiming in late), but that was a topic I was about to broach... (if there was a prior discussion, mind pointing me to the thread?)
I was going to propose considering adding kde-runtime in WG tech-specs. This will essentially ensure that, well, core runtime bits are already part of the expected install set.
If there are objections, I'd like to hear them, in case it's something that can be resolved through compromise and packaging modifications.
I would assume one big issue is size, which is understandable. Efforts can definitely be made to split out non-essential functionality to reduce dependencies and size impact. That will only go so far of course, underlying stuff like qt/oxygen-icons can't be cut.
-- Rex
Dne 5.3.2014 18:58, Adam Williamson napsal(a):
On Wed, 2014-03-05 at 09:29 -0500, Josh Boyer wrote:
Yep, and that probably leads back to the question if we want or don't want to provide more products than we currently have so far (three). And to be honest, I like idea having more products but with very high bar (that should apply for all products) - to avoid a situation of having dozens of products but still be inclusive.
All of your subsequent email is perfectly fine, but this thread is focused on what we want to do with KDE in the Workstation product. If you want to advocate for a KDE product, please take that discussion to the devel list. I don't want to derail this conversation, and this list isn't the right place to discuss what you're suggesting.
Well. This thread was sparked partly by my FESCo ticket, which is written from a project-wide perspective, not a product/WG-specific one. The project as a whole has skin in this game: it is important to the project how the project chooses to deliver KDE, and if the answer is 'in more than one way', the relative importance of those ways. The Workstation WG is free to integrate KDE into their product in any appropriate way, of course, but I think the project *as a whole* has a say in whether that's our primary method of distributing KDE - how heavily it's promoted, how hard it's tested, and whether it blocks our releases. In that context, the existence or otherwise and relative importance of a KDE spin is obviously significant.
We (the KDE SIG) will be discussing and planning how to achieve that in the following few days, we'll keep the relevant parties informed. At the same time, we are prepared to cooperate with the Workstation WG, should they want to integrate (parts of) KDE technologies into their product.
Thread closed here :)
Rex Dieter wrote:
Jaroslav Reznik wrote:
I heard, there were objections to have kdelibs in the tech specs. With move to KDE Frameworks 5, it could be raised again, as it's going to be more Qt 5 add on modules than old fashioned monolithic kdelibs.
(Sorry for chiming in late), but that was a topic I was about to broach... (if there was a prior discussion, mind pointing me to the thread?)
I was going to propose considering adding kde-runtime in WG tech-specs. This will essentially ensure that, well, core runtime bits are already part of the expected install set.
If there are objections, I'd like to hear them, in case it's something that can be resolved through compromise and packaging modifications.
I would assume one big issue is size, which is understandable. Efforts can definitely be made to split out non-essential functionality to reduce dependencies and size impact. That will only go so far of course, underlying stuff like qt/oxygen-icons can't be cut.
ping, I'm interested in helping work on this, but need help/feedback.
-- Rex
On Tue, 2014-03-11 at 09:28 -0500, Rex Dieter wrote:
ping, I'm interested in helping work on this, but need help/feedback.
Hi Rex, sorry for not getting back to this sooner. Its on my list for this week to look in more detail at the pros (or cons ?) of including kde-libs - I hope to get to it after Wednesday.
Size is one of the concerns that I've heard, but also a question whether we would end up with 'two of everything' - 2 multimedia apis (gstreamer vs phonon), 2 display configuration apis, etc. Sadly, I don't have very deep knowledge about the qt side of the world, so you could probably help me out there. How do things work nowadays: - Is qt5 more or less a complete platform ? - Are kdelibs mainly used by core kde apps, while 3rd party qt apps generally just use qt ?
Matthias
----- Original Message -----
On Tue, 2014-03-11 at 09:28 -0500, Rex Dieter wrote:
ping, I'm interested in helping work on this, but need help/feedback.
Hi Rex, sorry for not getting back to this sooner. Its on my list for this week to look in more detail at the pros (or cons ?) of including kde-libs - I hope to get to it after Wednesday.
Hi Matthias, see my answers - sometimes inline, sometimes not - I hope it's still readable :).
but also a question whether we would end up with 'two of everything' - 2 multimedia apis (gstreamer vs phonon), 2 display configuration apis, etc. Sadly, I don't have very deep knowledge about the qt side of the world, so you could probably help me out there. How do things work nowadays:
To make it even more complicated and confusing ;-) - the whole situation is going to change a lot with upcoming KDE Frameworks 5 (KF5) and Qt 5 but I hope it will answer all of yours questions. The main idea behind Qt 5 and KF5 is modularity (smaller footprint for mobile devices) but also with open nature of Qt Project, a lot of stuff from former kdelibs has been upstreamed to Qt 5. KDE libraries/runtime were split into so called tiers with the goal to provide these modules to other 3rd party Qt developers without the need to bring the whole KDE as it was in time of KDE 4.
Size is one of the concerns that I've heard,
With KDE Frameworks 5 the size issue could be solved by picking up the right modules from the tier that could be useful for wider range of users and developers.
For two of everything - usually as backends, WS libraries and frameworks are used (GStreamer for Phonon, UDisks for Solid etc.) and these APIs are than just more convenient APIs for Qt/KDE developers without the need to fight GLib/DBus integration and higher level than underlying APIs. We try to make sure we use same technology as current Desktop spin. And even getting closer for example with recent rewrite of network management using nm-qt directly.
- Is qt5 more or less a complete platform ?
Qt 5 is more or less complete platform, with addition of KF5 bits it's even more. But as I said - it's pretty much modular these days. You can just use Qt Base and write console apps without the need to have GUI libs available at all.
- Are kdelibs mainly used by core kde apps, while 3rd party qt apps
generally just use qt ?
That's the idea behind KF5 it's going to be used by 3rd party at apps developer as I wrote above.
Also the KDE libs and applications are now being decoupled, so it gives as some flexbility.
What I'd propose (it's my personal opinion, not KDE SIG) - let's start with KF 5 - it should be "stable" by F21 release. There's some possibility some apps would be already ported to Qt 5/KF5 and we will start from clean ground. It will be useful for developers beyond old KDE lands...
For everyone who would like to install KDE 4 desktop or KDE 4 apps, deps would have to be obtained from repository. But even Plasma 2 desktop is shaping up but I don't think it would be ready for general consumption by the time of F21.
Jaroslav
Matthias
-- desktop mailing list desktop@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/desktop
Jaroslav Reznik wrote:
What I'd propose (it's my personal opinion, not KDE SIG) - let's start with KF 5 - it should be "stable" by F21 release. There's some possibility some apps would be already ported to Qt 5/KF5 and we will start from clean ground. It will be useful for developers beyond old KDE lands...
For everyone who would like to install KDE 4 desktop or KDE 4 apps, deps would have to be obtained from repository.
There's currently a large collection of kde4 applications to choose from. The Qt5/KF5 set is (and will continue to be for some time) very limited and small. So, I think your suggestion to focus on KF5 has the priority backwards.
That said, similar to how both gtk2 and gtk3 are included, so should both qt4 and qt5... eventually (and ideally). I just think the starting point should be where there is the most benefit *now* (qt4/kde4).
-- Rex
Rex Dieter wrote:
That said, similar to how both gtk2 and gtk3 are included, so should both qt4 and qt5... eventually (and ideally). I just think the starting point should be where there is the most benefit *now* (kde4).
Correction, qt4/qt5 are both already included, I meant to say kde4/kf5 core/runtime support instead.
-- Rex
On Wed, 2014-03-12 at 07:52 -0500, Rex Dieter wrote:
That said, similar to how both gtk2 and gtk3 are included, so should both qt4 and qt5... eventually (and ideally). I just think the starting point should be where there is the most benefit *now* (qt4/kde4).
I did put both qt4 and qt5 on the package list. This discussion is largely around figuring out what we need to include beyond that to make most qt apps work.
On 03/12/2014 02:34 PM, Matthias Clasen wrote:
On Wed, 2014-03-12 at 07:52 -0500, Rex Dieter wrote:
That said, similar to how both gtk2 and gtk3 are included, so should both qt4 and qt5... eventually (and ideally). I just think the starting point should be where there is the most benefit *now* (qt4/kde4).
I did put both qt4 and qt5 on the package list. This discussion is largely around figuring out what we need to include beyond that to make most qt apps work.
Including qt4 and qt5 in Workstation is somewhat commiting us to API/ABI stability with these libraries, giving 3rd parties the signal that they can rely on these being always available. These kinds of promises, even if unspoken, make it difficult to remove them in the future.
Having said that, I think it's fine to include both gtk2/gtk3 and qt4/qt5 -- these are likely to be around for quite some time and apps are going to rely on them in the future.
I am not sure the same makes sense for kdelibs4/kdelibs5. If I understand the situation right, the former is likely to pretty much disappear, or at least fall into disuse when KDE gets ported to kdelibs5.
In particular, what makes the situation different here is that kdelibs is mostly used in KDE apps, which are likely to be ported fast. Qt on the other hand, is used by lots of 3rd party developers and I would assume it takes quite some time for them to switch to new versions.
My suggestion would be to include: - qt4 - qt5
and skip kdelibs for now. Once KDE5 is in Fedora, add kdelibs5 as well.
Matthias Clasen wrote:
This discussion is largely around figuring out what we need to include beyond that to make most qt apps work.
OK, baby steps for qt apps, additional items to consider... media support (highly recommended): +phonon +phonon-qt5
python bindings: (mentioning for completeness, I currently doubt adding these are of high enough value): +PyQt4 +python3-PyQt4 +python-qt5 +python3-qt5
-- Rex
Kalev Lember wrote:
I am not sure the same makes sense for kdelibs4/kdelibs5. If I understand the situation right, the former is likely to pretty much disappear, or at least fall into disuse when KDE gets ported to kdelibs5.
kdelibs4 will likely be around (and supported) at least until late 2015 (based on the claims that kde-workspace-4.11.x LTS being supported for 2 years, whose initial release was Aug 2013).
-- Rex
On Wed, Mar 12, 2014 at 09:01:52AM -0500, Rex Dieter wrote:
Matthias Clasen wrote:
This discussion is largely around figuring out what we need to include beyond that to make most qt apps work.
OK, baby steps for qt apps, additional items to consider... media support (highly recommended): +phonon +phonon-qt5
I'm assuming phonon-backend-gstreamer by default?
Matthew Garrett wrote:
On Wed, Mar 12, 2014 at 09:01:52AM -0500, Rex Dieter wrote:
Matthias Clasen wrote:
This discussion is largely around figuring out what we need to include beyond that to make most qt apps work.
OK, baby steps for qt apps, additional items to consider... media support (highly recommended): +phonon +phonon-qt5
I'm assuming phonon-backend-gstreamer by default?
Yes, so I guess we should explicitly include that too: +phonon-backend-gstreamer +phonon-qt5-backend-gstreamer
note this is still currently based on gstreamer-0.10. There is development underway to support gstreamer-1.x, and I'd say there's a good chance for that to land in time to be testable for f21.
-- Rex
----- Original Message -----
Kalev Lember wrote:
I am not sure the same makes sense for kdelibs4/kdelibs5. If I understand the situation right, the former is likely to pretty much disappear, or at least fall into disuse when KDE gets ported to kdelibs5.
kdelibs4 will likely be around (and supported) at least until late 2015 (based on the claims that kde-workspace-4.11.x LTS being supported for 2 years, whose initial release was Aug 2013).
I talked with Rex yesterday and his point is correct - at least for now, we should go with 4 too. But for limited time - I don't think it makes sense to stay in platform forever. It will also allow a better integration of KDE applications into Workstation Product to fulfil the idea of technology specific product but offer best applications from not only both G/K worlds - even the most will come from this world.
Jaroslav
-- Rex
-- desktop mailing list desktop@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/desktop
desktop@lists.stg.fedoraproject.org
-
Adam Williamson -
Christian Schaller -
Dan Mashal -
Jaroslav Reznik -
Josh Boyer -
Kalev Lember -
Lukáš Tinkl -
Matthew Garrett -
Matthias Clasen -
Rex Dieter -
Richard Hughes -
Stephen Gallagher