First release of KDE Frameworks 5

[Posted July 7, 2014 by ris]

The KDE Community has announced the release of KDE Frameworks 5.0. "Frameworks 5 is the next generation of KDE libraries, modularized and optimized for easy integration in Qt applications. The Frameworks offer a wide variety of commonly needed functionality in mature, peer reviewed and well tested libraries with friendly licensing terms. There are over 50 different Frameworks as part of this release providing solutions including hardware integration, file format support, additional widgets, plotting functions, spell checking and more. Many of the Frameworks are cross platform and have minimal or no extra dependencies making them easy to build and add to any Qt application."

First release of KDE Frameworks 5

Posted Jul 8, 2014 19:09 UTC (Tue) by fandingo (guest, #67019) [Link] (18 responses)

As an outsider, could someone explain this focus on turning KDE into frameworks? I understand and appreciate the benefits that it can have for developers (independent release schedules, better compatibility, more widespread use of KDE libraries, etc.). I'm just not seeing what the benefit to a user is. The screenshots and articles that I've read about KDE 5 is that it's the application and workspace functionality isn't changing for users.

My fear is that KDE has undergone a considerable amount of work in the past 3 years to improve plumbing that isn't directly benefitting users. The developer benefits that exist aren't obviously going to help users.

I'm not saying that KDE is wrong. I'm just trying to understand the focus on turning everything into independent libraries and the improvements that users will see.

First release of KDE Frameworks 5

Posted Jul 8, 2014 19:33 UTC (Tue) by roblucid (guest, #48964) [Link] (6 responses)

Do you remember KDE4 development and initial releases? The key benefit to end user of this plan is that they don't break YOUR desktop. HLater, you get something stable on more platforms, or with new features and ideas from a larger pool of developers.

When they tried to port to Qt4, simultaneously creating an experimental new modern desktop, have ported applications all to be released in one lump. It was a PR disaster.

Conclusion was, they had to decouple things, so the user base who just need a stable desktop to work on, aren't disturbed or distracted. New features need to be exposed application developers first who are better placed to trouble shoot breakage. Secondly, now Qt is fully OSS, by releasing KDE core into the ecosystem and contributing code, they increase pool of developers to include mobile/windows ones.

First release of KDE Frameworks 5

Posted Jul 8, 2014 20:42 UTC (Tue) by DionNestor (guest, #97781) [Link] (5 responses)

>When they tried to port to Qt4, simultaneously creating an experimental new modern desktop, have ported applications all to be released in one lump. It was a PR disaster.

True. But the real disaster was not the bugs. The lack of KDE communicating the sad state of KDE 4.0 was the worst part. The release announcement had no warning. Improv is the same PR disaster now. Frameworks5 might fare better though.

First release of KDE Frameworks 5

Posted Jul 9, 2014 12:46 UTC (Wed) by sebas (guest, #51660) [Link]

"KDE 4.0" was exactly the problem. The codebase had scaled beyond what could be stabilized in one major cycle, there are multiple layers which depend on each other: One needs to make base libraries work before being able to work on apps using them. Also within apps, there were huge differences in maturity: Some games were perfectly stable way before 4.0 was released, the desktop was completely new and barely usable. The development model simply didn't scale, so these parts needed to be untangled.

Apart from that, more modularity has always been a wish of many developers and users. Many 3rd party app developers wanted to use code from kdelibs, but didn't want to link to all of it just for a small 'sublibrary'. This has stood in the way of adopting kdelibs code in a whole bunch of apps, and forced developers to do all-or-nothing decisions. With a focus on more form-factors, this footprint has become even more important, and that has been addressed in Frameworks.

To me, this might not have a direct benefit to the user, but it has important implications to developers, deployers and apps. We've also had reports that startup of apps has become a bit faster, but I haven't measured that myself (got the same impression though), so unless someone independent (and non-You-Know-Who ;)) benchmarks it, I won't claim that as a user-visible benefit.

Sometime, one has to take a step back and invest in the foundations to be able to keep moving forward.

All that said, the new Plasma is almost ready, and there you'll get a lot of user-visible changes (and it's based on Frameworks 5).

First release of KDE Frameworks 5

Posted Jul 9, 2014 13:38 UTC (Wed) by Sho (subscriber, #8956) [Link] (3 responses)

> Improv is the same PR disaster now.

Note that Improv communication isn't handled by the kde-promo group (Improv isn't technically a KDE project but a downstream of KDE bits, though of course there certainly was people overlap, just like between KDE and other downstreams).

First release of KDE Frameworks 5

Posted Jul 9, 2014 13:44 UTC (Wed) by rahulsundaram (subscriber, #21946) [Link] (2 responses)

It is unclear to me what the extend of overlap is and what makes it an unofficial project. A communication strategy to make it more clear would help.

First release of KDE Frameworks 5

Posted Jul 9, 2014 14:11 UTC (Wed) by krake (guest, #55996) [Link]

It is not unofficial either, it is just not a KDE project.

It is hardware by a hardware manufacturer, said to be capable of running KDE software.

People working on KDE work on all kinds of things outside their contributions to KDE.

First release of KDE Frameworks 5

Posted Jul 11, 2014 23:55 UTC (Fri) by Sho (subscriber, #8956) [Link]

For the record, KDE projects are those that abide by the "Commitments" listed on this website: http://manifesto.kde.org/

The KDE Incubator program helps projects come into compliance with them.

Improv never applied for that, and I'd wager probably also didn't want to (since the people behind it likely knew they were taking a risk and wanted to isolate KDE from that).

The Improv folks certainly did a lot of nice work upstream participating in KDE and contributing there, but the Improv was hosted by them separately, the finances were handled by them, interacting with customers was handled by them, and so on. Bottom line, it was an attempt to launch a product that was using KDE bits, and those behind it interacted well with the community. Shame it didn't go better.

First release of KDE Frameworks 5

Posted Jul 8, 2014 20:11 UTC (Tue) by DionNestor (guest, #97781) [Link] (6 responses)

>As an outsider, could someone explain this focus on turning KDE into frameworks?

KDE does not have the developer resources it used to. They went from desktop environment to meta. Frameworks5 is just a new modularized extra to Qt. KDE mentions Qt 13 times in the announcement including one picture putting Qt ahead and above KDE.

This might add some concerns about corporate independence.

First release of KDE Frameworks 5

Posted Jul 8, 2014 22:41 UTC (Tue) by smoogen (subscriber, #97) [Link] (3 responses)

I think you are confusing symptoms (corporate dependance) and causes (lack of developers). Where did all the KDE developers go? They didn't go to GNOME or any other Linux desktop (any more than the disaffected GNOME3 developers went to KDE).

Did they go to Apple systems and Appstores? I don't know.. but I think it is giant missing elephant in the room that needs to be found.

First release of KDE Frameworks 5

Posted Jul 9, 2014 4:36 UTC (Wed) by DionNestor (guest, #97781) [Link] (2 responses)

> I think you are confusing symptoms (corporate dependance) and causes (lack of developers). Where did all the KDE developers go?

I guess many developers went on with life. Today software development is so much more diverse. Being meta Qt software is very different to offering a full desktop environment. The latest blog about where KDE is going shows this very clearly.
http://lwn.net/Articles/604134/

I think it makes sense for KDE to transform. We got enough Linux desktops anyway. The Qt world is bigger than the Linux desktop.

First release of KDE Frameworks 5

Posted Jul 9, 2014 16:16 UTC (Wed) by drag (guest, #31333) [Link] (1 responses)

> The Qt world is bigger than the Linux desktop.

Probably not for long then. I really don't see a 'cross platform toolkit' as a really compelling thing in the future.

First release of KDE Frameworks 5

Posted Jul 9, 2014 18:20 UTC (Wed) by rleigh (guest, #14622) [Link]

What do you see as compelling if not this?

If you prefer that using a toolkit implies tying oneself to a specific desktop/OS/device, that greatly limits who will be able to use it. I can certainly appreciate this sometimes can make sense, but in the general case it's often unjustifiable.

First release of KDE Frameworks 5

Posted Jul 8, 2014 23:16 UTC (Tue) by HelloWorld (guest, #56129) [Link]

> KDE does not have the developer resources it used to.
[citation needed]

First release of KDE Frameworks 5

Posted Jul 11, 2014 23:57 UTC (Fri) by Sho (subscriber, #8956) [Link]

Note that KDE also contributed thousands of lines of code to Qt 5. This whole notion of "extra" is silly because it implies an us-vs-them divide that we all worked hard to get past. Qt is a working open source project now, and the KDE community is one of its biggest stakeholders and contributors -- heck, we hosted the last Qt Contributors Summit at a KDE conference.

First release of KDE Frameworks 5

Posted Jul 9, 2014 12:37 UTC (Wed) by krake (guest, #55996) [Link] (3 responses)

> As an outsider, could someone explain this focus on turning KDE into frameworks?

It is not about turning KDE (assuming you mean the desktop product with that) into frameworks.

The KDE Frameworks 5 effort is about taking KDE's extensive libraries and carefully examining, defining and shaping each library's boundaries.

The previous incarnation of the KDE libraries product, usually referred to as KDE platform, looked to outsides like a single library or at least like a single package containing all these libraries.

The Frameworks product makes each library visible with its specific boundaries, i.e. allows developers looking for certain functionality to see which library would have that functionality and which dependenices that requires.

> I'm just not seeing what the benefit to a user is

Well, the users of libraries are application developers. The benefit for them is that they get a better view of which library does that, which library they might not need, better tested and more actively maintained libraries.

Applications using the current libraries will get the same functionality, potentially with better package dependencies that now.

Applications currently not using KDE's libraries can potentially extend their feature set or replace custom solutions with sharedly maintained ones.

First release of KDE Frameworks 5

Posted Jul 9, 2014 12:56 UTC (Wed) by sebas (guest, #51660) [Link] (1 responses)

Expanding on that, briefly:

What you call "KDE" is actually three things:

- a set of libraries
- a desktop environment
- a bunch of applications

These three products (or sets of products) are now released independently. What you see here is the continuation of the development framework.

It's not about "turning KDE into", it's about splitting up the products of the KDE community along meaningful borders. We have started on this by defining these products independently (and by defining "KDE" as the community, the team who create it). The libraries are nowadays called "Frameworks", the desktop environment "Plasma".

The applications already have their individual name and brands, but might very well come in a collated fashion, in sets of apps that work well together. We haven't settled on a final distinction between sets of apps yet, but that will surely be one important question for the near future -- discussions which apps should be shipped together as logical packages has started, but not concluded yet. I could, for example, imagine to have a "base suite" with file manager, image viewer, a PIM suite (we have that), a creative suite (sorry Adobe ;)), a multimedia suite, etc., all with apps that together make up nice workflows. This kind of view will also have effects on how apps are developed and laid out, so it potentially can have huge impact in the long run.

First release of KDE Frameworks 5

Posted Jul 10, 2014 17:03 UTC (Thu) by roblucid (guest, #48964) [Link]

This excellent explanation left out, reference manual documentation and user guides which focus on key concepts. It's a really important part of making a project really accessible and usable.

Being able to choose, just what you need, helps reduce the "full fat" image, Desktop Environments suffer from.

First release of KDE Frameworks 5

Posted Jul 10, 2014 17:09 UTC (Thu) by roblucid (guest, #48964) [Link]

Shared maintenance isn't only benefit of code re-use. It tends to improve quality over time, as it is applied in new situations or ported, and is worked on by a wider vareity of people with exposure to wider range of techniques. Good example code, helps write better new project code applied to different problems, as it reduces time spent on solved problems with clear neat solutions and examples.