a relaxing friday

I took some time this evening to relax in a quiet, empty house by plonking away at the Plasma Active Alarms app UI. I'm still not 100% satisfied yet, but it's getting a lot closer.

Alarms plasmoid in Application form factor with touch components

Under the hood there are a lot of really interesting things going on. For instance, it's implemented as a Plasmoid that supports the Application form factor; the first Plasmoid to do so, in fact. What this means is that you can pop it on your desktop, in a panel or run it as an application and it adapts its UI appropriately. This is a new feature for 4.10 and I wonder what other Plasmoids may make similar leaps. Alarms also works very nicely in a "single column" form factor, smartphone style, when the horizontal space is constricted.

It uses Akonadi's spiffy KAlarm integration, so the alarm infrastructure was done for us by the KDE PIM team. This grants the Alarms app integration with all the other KDE PIM Stuff(tm) for free, too. Communication with Akonadi is done via a typical DataEngine and Service combo (org.kde.alarms). Due to this, without having to extend Plasma's C++ or QML API one bit, it was instantly and easily accessible in the QML as a model.

Still, there are caveats: The typography still sucks as we have not yet implemented the new guidelines in the components, something I want to see happen for Active 4. At least it is using all the "right" components so it can benefit as soon as the typography stuff slots into place. The plus icon looks dated by now (I'd guess it is one of the earlier Oxygen icons), and I'm not convinced with position of confirmation buttons in editor or the create entry ... the usual "must be perfect" itches.

But it's getting there ... at which point it'll be ready for features again such as more flexible repetition, selecting the audio file to play, etc. 

quicky update on window grouping

I posted a couple of blog entries the other week about how to show grouping in the task bar. The feedback was fantastic and we now have a "winner", at least for 4.10.

Many noted that knowing there is a group of windows just by glancing at the task bar is important to them. Very, very few find value in the little number, however. So putting it all together, in 4.10 a small arrow will be shown drawn over the window icon when there is a group, as is seen in many other dock apps. This preserves the most useful information while also limiting visual noise, creating a more compact layout in the window listing and finding consistency with other implementations out there in the wild.

As a bonus, the arrow is no longer the (in my opinion) ugly black thing that was specific to the window list but the standard arrow used throughout the rest of the Plasma workspace interface.

Thanks to everyone for your feedback!

Bonus time! ;) A curious little walk down memory lane: in the later 3.x releases of kicker (I forget if it  was 3.4 or not; but around that time, I believe) I implemented a new look for the tasks widget which was made configurable via the task applet's configuration dialog. The old look was kept as "classic" and the new look was named "elegant". When Plasma Desktop 4.0 was released, the tasks widget was rather different and looked a lot more like the "classic" style, but I kept to my commitment of deferring these things to the people doing the graphics, as long as they did not impact usability (which would trigger a discussion to find a consensus-base solution). With 4.10 we are now back to something much like the default "elegant" layout in later 3.x tasks applets. The grouping element is now nicer, but otherwise .. it's "elegant" from the old days. It's funny how sometimes things go full cycle like that. I'm just enjoying my nice, calm panel with fewer lines and borders.

Current master, which will eventually be released with Plasma Desktop 4.10

You may notice the Share Like Connect widget next to the activities button. That will not be a part of the default 4.10 setup (likely in 4.11, however), but you an already install it from the share-like-connect repository.

help test task grouping experiments

In my last blog entry I shared some experiments with the tasks widget I did this morning. A good number of people have offered feedback so far, which is great. It can be quite difficult to guess at the usability impact of something like this with just screenshots, so let's take it to the next level, shall we?


I have uploaded a tarball of a modified tasks widget here and I'd love for you to try it out for at least a full day.

You will need cmake, a C++ compiler and the kdelibs and kde-workspace headers installed.  On OpenSuse, the kde-workspace headers are in the kdebase4-workspace-devel package, for instance, so a simple "zypper install cmake kdebase4-workspace-devel" should hopefully get you most of the way there.

Once you have downloaded the tarball, extract it to somewhere on disk and then create a build directory inside of the resulting directory. From in the build directory do something like: "cmake ../ -DCMAKE_INSTALL_PREFIX=`kde4-config --prefix`".

If all goes smoothly you can do a `make && sudo make install && kbuildsycoca4` type thing and you'll now have a new plasmoid called "Tasks Experiment" in the UI and org.kde.tasksexp behind the scenes. It won't trample on your existing tasks plasmoid or anything.

From there, remove the tasks widget from your panel and in its place put a tasksexp widget. You can even do this from a command line with something like this:

> kquitapp plasma-desktop
> perl -pi -e "s,plugin=tasks,plugin=org.kde.tasksexp,g" `kde4-config --localprefix`/share/config/plasma-desktop-appletsrc
> plasma-desktop

Voila.

You can also try it with the "stack icons" grouping by opening up the abstracttaskitem.cpp file and removing the two /s on line 71 so that GROUP_INDICATOR is defined. Then go back into the build directory and do the make / make install dance again, kquitapp plasma-desktop and start it up again.

After a day or two of using it, let me know how it works, or does not work as the case may be, for you.

 

mark grouped windows as such or not?

In the default tasks widget (the one that shows your currently open windows) if there is more than one window grouped together it shows an arrow and the number of windows in that group. This results in a panel looking something like this:

Tasks widget with group status information

Great, so I can see I have two konsoles and three kontact windows open. However, today I challenged myself to defend why that information is there. Is it actually useful? Does it matter if there are 2 or 3 windows in the group? Is it important to be able to visually identify it as a group?

With or without this additional information, the number of clicks to get to a window in a group is the same. Click once on the group, and then again in the resulting popup. If compositing is active, you can also just hover the group and then click on a window thumbnail.

I could not think of a single time when the extra visual information of the arrow and number ever provided value to me, and I struggled to come up with a use case where it might. So I patched out the group status information just to see what it looks like and to see if it would negatively impact how it works. It took all of probably two minutes to do and I've been using the tasks widget since without grouping status information, as can be seen in the shot below:

Tasks widget without group status information

There is more room now for window titles which is nice, and there is less information being presented resulting in a more elegant look.

Visual simplicity, however, should usually not trump utility, however, so I'm trying to come up with uses for the group status information. Just because I can't think of one off the top of my head does not mean that a use case does not exist. So I thought I would ask you! Do you find this group status information useful? If so, why? If not, would you prefer it not to be there? Leave your answer in the comments section below!

Even if you don't have input to share, you at least get a glimpse into the sort of banal questions we get to ponder when working on Plasma. Not everything is about exciting shiny new toys after all. :)

Update: Decided to try another possibility: show grouping by painting the icon twice, once full color and a second time faded. Keeps the visibility of "this is a group" without taking up more space and with less visual noise. Not convinced by this, but was something that occurred to me as a possibility. :)

Using a visual "stacking" of the icon to denote a group.

p.s. Please ignore the misalignment of the system tray in the screenshots. It's a known bug in the QML UI that is new for 4.10. It will be fixed before release.

p.p.s. Please also ignore that the task icons are a different size than the other icons on the panel. I'm considering possible improvements for that too as all icons being the same consistent size, at least when there is just one row in the tasks widget, would be rather spiffy.

p.p.p.s. I am not interested in suggestions to replace the tasks plasmoid with some other plasmoid in the default layout. That is a separate discussion, and I'd like to focus on this one specific issue here.

the sun is a billion little spotlights

The sun is a seething sphere of plasma consuming approximately 600 000 000 tons of hydrogen per second,  lighting our solar system with its raging burn.  Even at 150 000 000 kilometers away from the sun, our little blue planet receives 1 800 000 0000 000 000 000 Joules of energy from it every second, or over 10 000 times the energy humans consume. This immense amount of energy, spread across the Earth, allows all the weather, all the life and so much of the beauty we see around us.

You can focus the energy of a square meter of sunlight on earth with a Fresnel lens and this will provide over 1000 Watts of energy in a very small space. This is likely to hurt, or even extinguish, any living thing that finds itself in the way of this energy. Which is why it is pretty handy that the energy from the sun does not simply come out in a single beam of light, striking the earth in one place. Even with a handful of such sun powered spotlights, there would be a few areas with hellishly intense levels of energy while the rest of our Earth (assuming it would exist at all) would lay cold and desolate.

This is why I like to think of the sun as a billion little spotlights, each shining on some patch of our earth. Every plant, every animal big and small, gets an opportunity to have a little light shone upon then. Enough to power life, but also not so much that it is destroyed. Perhaps more fancifully we could think of daylight as the combined effect of all those spotlights. :)

The free software community (you knew it was going there, right? ;) is a seething community of millions of humans generating amazing amounts of creative work. This is an amazing source of attention energy with people looking for answers and answering questions and creating new questions for people to find answers to. This energy tends to get hyper-focused at times, robbing some areas of our little planet of code of needed sunlight and bathing others in cancer-causing dosages of it. I think we can be like the sun and spread it out a bit more. We can cast a million smaller spotlights, and while that may make the light a little less exciting, the results will be a rather more lush environment.


Armageddon levels of energy may be exciting, but I'd rather spend my days in a field full of flowers.

On Google+ today I read this from John Layt:

Great weekend in Brno at the Linux Color Management Hackfest, a mix of Oyranos, colord, KDE, Gnome, openSuse, Red Hat, and Ubuntu people all just getting on with figuring out how to get things done.  There wasn't even a single heated argument, so don't expect it to hit the headlines any time soon :-)  We really need more of these type of events to build communities around the common desktop infrastructure.  Now to try organise and write-up my thoughts for posting to kde-core-devel and qt-development, could be an epic!

Awesome! But ... "there wasn't even a single heated argument, so don't expect it to hit the headlines any time soon"? Really? Can any of us be proud that this is where we are, that unless arguments surpass a certain (arbitrary) threshold of controversy we don't expect it to get attention? Is this not the best way to breed bad behavior and rob the spirits of those who put in honest efforts?

So let's shine a million spotlights on the things we're achieving, like this color management hackfest. Let's take a low-powered moment to celebrate the working-togetherness and positive results of it. Let's see that on the front page of Slashdot; let's read about it (and a dozen other daily events) on our blog planets. We'll have less time and energy left over to fanboi obsessively over other things that are, relatively, less consequential. And that's OK. It would be a more complex story to follow, just as it is harder to pick out an individual plant in a dense youthful forest, but it would be thick with life.

In a technology industry where the dominant pattern right now is to hyperfocus on a few products from a few companies driven by a few myth-level superstar heroes (who, like all myths, are mostly not real), this would be an amazing source of advantage for Free software in the same way as the sun's light being spread across the whole of our planet is to biological life.

It may seem that there is a fly in the ointment here: this may not scale since people who wish to just sample what's going on, rather than devote their life to keeping up with Free software, only want to hear about the "big" events and high-level summaries. Fortunately for us, we have news organs such as Linux Weekly News that can do just that. Interestingly, I have noticed a lack of obsessive hyperfocus dysfunction in such "we summarize the news for you" outlets, particularly when the community being covered is diverse and doesn't hyperfocus. In the grass-roots Free software community, we need to set the (plurality of) topics in the community rather than allow ourselves to pulled about by the gravity of these "summary level" efforts. We need to be (positive) sources rather than reactive agents.

So let me put my spotlight where my mouth is: To everyone who was at the Linux Color Management Hackfest: you are my inspiration for today. I hope you had an awesome time and I can't wait to use (and tinker with, of course ;) the results when they find their way into a nearby source code repo. :)

cults of personality redux

In a recent blog post, I slammed cults of personality in Free software communities. Some noted in the comments that this was not the only challenge we faced, and I completely agree. On the one hand, it's a bit of an odd observation to make: of course all complex results have complex sets of causal factors. Entire volumes have been written about this aspect of complex challenges, and a thread I've noticed in a number of pieces I've read is that the shear number of causal factors makes it hard for people to untangle and overcome the challenges presented. It's like we become distracted by too many topics and forget that you eat an elephant one bite at a time and not all at once; that it is OK to examine and address issues in a piecemeal fashion.

Others noted that there are some good affects that come from these cults of personality. This is also true. But it's sort of like saying, "Since I put $100 in the bank today, I will have $100 to spend." That may be true, but if you already spent $500 using your credit card .. no, you don't have $100 to spend. You owe $400. This is simple math that most people get intuitively, but when we apply it to systems analysis it often gets missed. Most things people do have some advantage (locally, individually, in the here-and-now, etc), and that is usually how they get entrenched in the first place. Full accounting, which means looking at the broad spectrum of results, is required to come to a full and proper sum, however. Some benefits are not good enough when there are large amounts of documented negatives.

Then there were people who said that they agreed with Linus. This is accurate, also. However, when looking at systems issues (which is what a cult of personality is: an attribute of a community, which is a system of social actors, in this case humans) the individual interaction is not the whole picture. I was certainly not asking if people agree with Linus, or if Linus is correct in any given statement. The issue is whether or not it is healthy to continuously and universally elevate a specific individual's opinions.

In case the answer is not clear to you, consider where this is actually practiced (dictatorial government, religion and pop culture media, to give three examples) and where it is not (the scientific method, participatory democracies) and then reflect on the empirical results each produces. One set puts humans Mars and works to ensure human rights are respected; the others give us things like tabloid magazines, inspires violent radicals and robs people of freedom in the name of expediency for the few.

I was really not interested in issues of personal agreement, but systemic affect. This is what I meant when I wrote, "I don't care what Linus says." I might agree with him, and yet my issue with putting that opinion on the front page of every F/OSS news site still stands.

Finally, some noted that there was perhaps a cult of personality around me and so I shouldn't throw stones, living as I am in a glass house. I truly hope no such thing actually exists, but if it did I would actively discourage it. I see myself as encouraging critical thought, attempting to inspire others to achieve more and organizing efforts where beneficial. In doing so, people may end up paying more attention to what I say than the average Free software contributor, but it is not (or so I hope) because of a blind belief in an abstract "Aaron" but due to the individual thoughts and experiences I share as I share them. I also hope that my attempts to spread awareness of the work of others, to highlight positive issues rather than only hipster-rant about negatives, etc. has a positive effect beyond my own standing in the community.

I see others in the Free software world doing this same kind of work as well, often better than I do it in fact, and that set of philosophers, organizers, leaders and generally motivated individuals are whom we ought to pay attention to in the areas where their efforts are applied. We would benefit by focusing some of the attention we lay on F/OSS "super stars" on those people.

That, really, was the entire plot point of that "cult of personality" blog post. :)

Ok, enough of that .. let's move on. (continued ...)

plasma sdk accreting

If you have used plasmoidviewer, plasamengineexplorer or plasmawallpaperviewer from past releases of KDE Workspaces while developing Plasma components, you may be surprised to find them gone in the upcoming 4.10 release.

Well, they aren't actually gone: they have been moved to the plasmate git repository. This is where the streamlined Plasma add-ons creator has lived in nearly perpetual alpha/beta for longer than I care to admit. It's finally in the release review process (aka "kdereview") and we decided to take this time to remove all the bits that are only interesting to developers from kde-workspace and put them in one place.

 This also opens the way for us to share more code between these developer tools, such as the Plasmoid previewing that is in both plasmoidviewer (obviously :) and Plasmate.

If you are concerned that you use plasmoidviewer to run individual Plasmoids in a window, do not despair! You ought to be using plasma-windowed which does a much better job of this and which has been available in kde-workspace for quite a few releases by now.

So now if you clone the kde:plasmate repository you will find all these developer tools in one convenient place. Enjoy :)

ending the cults of personality in free software

Free software has a history of creating and supporting cults of personality. Since it is a widespread human phenomenon, it is easy to understand how this happens. It is, however, unhelpful and destructive and we really ought to actively discourage it, starting by putting aside the current cults.

The most recent, though certainly not the only example, of this retrograde behavior was seen when Linus Torvalds started posting on Google+ about his recent sampling of various Free software desktop options. This made the rounds on various blogs, Slashdot, Linux Today, etc. Why do we care what Linus uses? Simply put: he's a super star in F/OSS and people hang on his every word as a result as if he's an oracle of all things technology. Spoiler alert: he's not.

His opinion on desktop software is as meaningful as his opinions on rocket ships, film production, oil recovery techniques, sociology, religious history, automobile engineering or any of the other topics he has no meaningful expertise in. Everyone will hold an opinion on things they use, such as cars or ships, and things they simply come into contact with, such as film and religion. Sharing our viewpoints is great and a necessary part of democratic discourse which can move societies along their evolutionary path. So opinions have value and can be hugely beneficial when mixed together with hundreds, thousands or even millions of other samples. The trick, however, is not letting those whose opinions are of no great value in terms of being based on greater understanding or access to relevant data become more important than other opinions in those discussions.

Linus is as likely to be "wrong" about desktop Linux as most others who use the desktop. I used quotation marks because he isn't really wrong at all: whatever his viewpoint de jour is reflects his viewpoint very clearly and may even be extrapolated to other people with similar needs and backgrounds. This should not be surprising. It's like saying Linus is as likely to be wrong about rocket ships as anyone else who is into rocket ships but not involved in their design and manufacture. Perhaps more so, even, if Linus isn't a big space enthusiast but only an occasional article reader and watcher of things NASA puts out press releases on.

When I first saw others commenting about his comments on Google+ about trying KDE Plasma, I shared the above sentiment, but in a more compact form. I don't care what Linus uses, at least not as an isolated data point: it is not useful information for one simple reason: it does not help us create the software the world needs.

Let's step to the side and consider this from a different angle: Imagine that someone made Linus' perfect desktop environment. Something that satisfied him entirely and which he could happily talk about whenever he felt like it. Would that environment be interesting and useful for the general public, or would it be something great for kernel developers and grumpy-heads like Linus? It could go either way, really, because (once again) the fact that Linus liked it would not be useful information when held in isolation by itself.

We could pick any other single individual on the planet who lacks the same amount of relevant experience and say the same thing. This is not a bash on Linus, it's an observation of something that Free software world does very poorly. It is not Linus' fault we pay more attention to him on such topics than we ought to; it's us who decide to pay attention to Linus on such topics. We can also decide not to. Even better, we could focus on amalgamations of well formed data sets and pay attention to those instead.

So why, then, do we pay so much attention and care at all what Linus says or thinks on these matters? There are very human reasons for why we do this, which the social scientists in the crowd can probably speak to rather better than I can, so that's a largely rhetorical question. I believe it is undeniable, however, that we risk the direction of Free software by building such cults of personality. It create incentives to acquiesce to what amounts to low-value personal opinion which leads to a meandering pattern of non-progress.

Of course, it goes even further than that. These cults of personality encourage others to mimic people like Linus when it comes to things like communication style. Due to the "we mimic which we admire" tendency in human, holding Linus aloft encourages people to adopt incorrigible asshat communication patterns online. Linus is not a role model for anything other than running very successful kernel (and similar) projects, yet he gets emulated by people who mistake the secret of his success with his communication style.

The cult of Linus is not the only center of a personality cult in Free software, to be certain. This episode was simply a good opportunity to say something about it. I could just as easily write a blog entry about the problems other cults of personality have caused in Free software. My years in the community have "blessed" me with a great big bag of personal experiences related to just about every single out-sized personality in Free software.

... which is the "big lesson" here: nobody is a messiah and we are all equally frail and imperfect. In more positive words: we are all equally infallible and perfect. So where can we derive leadership from, if not from luminescent personalities?

I don't believe Free software needs, or even benefits from, global leadership, anymore than the world needs dictatorships or a single global government. What Free software needs is local leadership, and those leaders need to have their influence limited (by a healthy lack of deification, among other things) in order create an environment in which it is valuable (and more acceptable) to reach out to others across domains.

Without this, we drift ineffectually. Free software was not able to truly take off on consumer devices until a corporate dictator stepped in, basically ignored everyone else, and pushed out Android. That in turn has become a primary source of GPL violations in today's world (I just found another one today, this time in a tablet being offered for sale by a very nice and reputable Free software supporting computer retailer) and is not an ecosystem with equal participation or open governance as hallmarks. We needed Android because we couldn't do it ourselves, but Android is also a huge weakness for Free software efforts and falls well short of what is realistically possible. We can't complain too loudly though, because we couldn't get our shit together to produce better results sooner.

We're still waiting for this savior on the desktop side. Some thought it would be Ubuntu, but what we got from Ubuntu, largely due to a "self appointed dictator for life" approach, is a fork of one of the major desktop environments and its community with a unhelpful sideshow of nearly endless backbiting between various Linux distributions, kernel developers and others with Canonical. All of this happened for understandable reasons, none of it was done without internally consistent logic, and yet it has not furthered anyone's ultimate goals for Free software success.

I don't know about you, but I am sick and tired of Free software being inefficient and self-destructive due to internal schism.

Ah, but we do have a positive example where Free software has done very well: the server and super computing. Why has Free software, driven by the open source development model, shone so brightly here? When we take a deep look at its history we may note the relative lack of personality cults and the amount of emphasis on data that matters (something that is easier to accomplish when one can use performance benchmarks or compliance suites as respected and reliable metrics). Certainly there are people who are respected, admired and influential in server-side Free software, but there has been a distinct lack of personality cult and a huge number of of cooperative (even in competition), technology-centric efforts. Samba springs to mind immediately for me with the likes of Jeremey Allison and John Terpstra, two people I highly respect but who don't have that "cult of personality" aura.

It's not all roses on the server side all the time, of course, and as a counter-example to Samba I'd probably point to MySQL and its travails over the years (from early non-Free software status, to odd GPL interpretations to corporate take over), which I believe directly correlates to its centralized control by charismatic leadership coupled with project fanboyism. You know, just like we see with desktop Linux.

To those who may agree but reply with "that's just how people work, not much we can do about it": well, I accept your POV and humbly disagree. We may not be able to change the world at large (at least, not over night), but we can positively affect the parts of the world we create ourselves. We create Free software together, and therefore we can create it in the image we want it to be.

As a participant co-creating this world, I have an impact on how it works and the culture that surrounds it .. as do you. We are not completely powerless to ourselves; we can make wise choices if we choose to. The progress of social justice in human history has shown that we are not slaves to our human foibles, though it can take time and effort to emancipate ourselves.

We can make the decision to free ourselves from cults of personality by taking our own advice: build decentralized structures of responsibility, influence and attention that reward things like cooperative effort. This will allow us to bring our efforts together into one combined and massive bonfire rather than thousand tiny embers fighting for fuel and oxygen such that they twinkle in but then almost inevitably back out again.

We need to realize that the behaviors that arise from our current culture do not further the goal we all share together of Free software success. If that is truly what we want, then we need to alter that which retards our progress towards it. We need to value Free software progress more than we value the reward of mutual fanboyism.

Put simply: if we want open client-side software to thrive we need to engage in a conscious improvement of the Free software culture and, just as importantly, leave behind those who choose not to. I won't try to fool anyone: this will take years to accomplish ... but we can start now.

Go ahead and say: "I don't care what Linus Torvalds uses on his computer." See how easy it can be? :)