DEC
20
2010

Introducing K16 and the Future of KDE

Where will KDE be in five years? To answer this question, we plan to bring together visionaries, strategists, planners, out-of-the-box-thinkers, realists, dreamers, doers, creators, leaders, coders from the KDE community and everybody else who is interested in discussing the future of KDE and picturing what it will be.

As a first step, we plan to hold a meeting under the name 'K16' to gather a first round of community people to come up with an idea where KDE will be in five years, in 2016. The goal is to formulate a common direction for the community, creating a shared vision that will drive the long-term development of KDE software.

The meeting will take place in early 2011 over an extended weekend. We'll provide a creative and productive atmosphere by having a heterogenous group of participants, inviting external moderators, having computer-free time, applying brainstorming and other creative techniques, and a location without distractions.

Join us

If you want to attend the meeting or provide input, please submit a one-page pitch of your idea where KDE will be in five years. The format and content is up to your imagination, be it a strategic text, a picture of KDE's Wikipedia article in 2016, a comic strip, or something completely different. The only condition is that it can be printed on a sheet of paper and pinned to a whiteboard. Please submit your K16 pitches to k16@kde.org no later than 31 December 2010. Please also state whether you would like to attend the meeting and when you are able to do so.

The K16 organization team (consisting of Sebastian Kügler, Dan Leinir Turthra Jensen, Riccardo Iaconelli, Stefan Werden, Jos Poortvliet, Artur Souza and Frank Karlitschek) will select up to 15 participants for the meeting based on the pitches, and invite them to the meeting. KDE e.V. will support the meeting by covering costs for travel and accommodation.

All results of the meeting will be published, and follow-up actions will be planned with the community.

We are looking forward to all your ideas. Don't wait, send your one-page pitch now to k16@kde.org. Share your vision of where KDE will be in 2016.

EDIT: the K16 team would like to add the following elaboration to the request for K16 pitches:

We would like to explain that we have some requirements for a 'vision' to be actually called a vision.

First of all, just stating what extra work is needed, especially when it comes to polishing what is there (no matter how important that is) is not enough. Yes, we all have many pet peeves and there are many areas where improvements are needed. But we asked for a 5 year vision.

Second, you need ideas on how to accomplish the plans within a Free Software community like KDE - and this is important. We are mostly volunteers, working on what we like or consider important. Whatever vision you present, it has to show how to get there and how to get people involved with it.

So in short, a vision should be ambitious - yet have some realism. Our long-term goals - create something so compelling it can overtake windows - should be the focus, not incremental improvements because that won't be enough in 5 years.

Comments

How about stabilizing and optimizing things instead of trying -once again- to revolutionize the desktop? You guys keep adding unnecessary things one on top of the other for no good reason.

For example the "activities". I don't know who will ever use them besides developers with OCD. Most KDE applications (like konsole, kate, konqueror, etc.) allow you to recursively divide the workspace, on top of that you have tabs, on top of the tabs multiple windows and on top of that even multiple desktops. Do we really need yet another layer on top of that? I know that the idea behind it is to "keep the state" but almost everyone use the same programs everyday so it doesn't really matter. For KDE 5 we may also have "Superactivities" (see the OCD problem).

And how about the obsession with metadata for everything? As long as the metadata isn't stored in the same file as the data itself, it is a dangerously bad idea. On many occasions the metadata is more important than the data itself, for example, the notes taken on a PDF document. This valuable data is stored in some obscure place inside the .kde4 directory. You cannot copy a file with its metadata using the command line, through FTP or a web server, or any other non-kde way. This is a trap, even under the very generous assumption that no bugs are ever going damage or corrupt this metadata.

The things that are really useful and matter, seem to be the ones receiving the less attention of all. Printing support sucks for anything non-trivial, Dolphin is buggy and slow, the kio-slaves used to be great, now they are also buggy and slow, kmail sucks, kopete is still stuck in the 1995 and okular is so obsessed with supporting all kind exotic formats that it sacrificed many important features a PDF reader should have (once again the OCD problem).

What you should do instead of dreaming about a psychedelic future is to go and hang Aaron for plasma and then care about the _present_ of KDE.


By trixl at Tue, 2010/12/21 - 9:53am

...suggesting that the focus for KDE in the next five years should be stability - no new features? - present it at K16 and win everyone over.

On to specifics:

So, if Activities are not useful for you, use only one. I don't see how they cause you a problem.

The Okular notes issue and why it's pretty complicated is explained in length in bug reports (and there is work ongoing).

Dolphin, here, is awesome, KMail is getting a rewrite ('sucks' is a bit non-specific anyway), kopete is old (still pretty good though) but there are new solutions coming for that kind of thing.

And lets not make it personal about Aaron, hey? I mean hang him for his karaoke singing, fair enough, but not for Plasma :-) You'd need to take out the rest of the team anyway - a lot of people like Plasma and are working on it.


By Stuart Jarvis at Tue, 2010/12/21 - 10:21am

>no new features?

I'm not against adding features to the individual applications, but I prefer applications that work well and fast, over those having many options that don't work.

KDE has too many programs that nobody cares about. The core functionality: a window and session manager, a terminal, a file manager, a messaging application, a pdf-reader, and mail client should be kept rock solid and fast because everyone is continuously using them.

I don't care if kpotatoman needs clustered SLI cards to work.


By trixl at Tue, 2010/12/21 - 11:02am

That was a bit of a silly reply from me in some ways.

But on a serious note - the meeting is about KDE in the next 5 years... Where should our focus be? Is there a need for lots of new things or should we work more on the polish and picking out a few key problem areas to work on. Can KDE do that?


By Stuart Jarvis at Tue, 2010/12/21 - 2:08pm

... but in fact you're right. KDE still has too many rough edges and the developers should put a stronger focus on those features most users actually need. Kopete is a very good example - I can't see any progress there and I don't think it is already perfect. Other examples are the KPackageKit (which is getting better but still lacks some features) and Konqueror/Rekonq (just does not have the same user experience as Firefox or Chrome).

So I think the first thing to do is to figure out some typical use cases, take some people without any KDE experience and see what problems they experience. If we want to convince people to switch to KDE, _everything_ has to be intuitive.


By Tom at Tue, 2010/12/21 - 10:50am

Agreed (except for the "hang" comment which is WAY WAY WAY out of line), everything should be intuitive (or at the very least easy to figure out).

But things should work!! To release unfinished/unpolished/buggy software (see my comments below on kmail, for example) in 2010 and later is just shooting oneself in the foot, since there are so many other options that work.


By Dulwithe Darkingham at Wed, 2010/12/22 - 7:35pm

i could agree with you but just on certain points,for example activities,honestly they are useless no one will ever use it,i am sure,of course you can always say that is a matter of taste but adding layer to layers is not a good thing,we should keep it simple and practical,for example i don't even use virtual desktops,even if i think is a nice idea but is more the effort taken by switching the whole desktop(for my little laptop at least) than opening/closing a window and we should keeping in mind that kwin now has tabbed windows so it is really becoming a jungle of tabs and windows and desktops and whatever,for kmail,you are just wrong i think kmail is great(it also won some comperition as the best linux mail client) is very well made and good looking,in the other hand you are right about kopete,i think is should be deprecated in a near future,KDE-Telephaty is coming and it will be great and freedesktop.org compilant so we don't need 100 avarage clients,but just one or two but very good ones;finally i agree when you say that's better to optimize what we have instead of try to revolutionize the desktop again,i don't comment on okular and dolphin becuse they are great piece of software and you should at least explain why dolpinh is slow or okular sucks(??!!),btw all kde software rocks,and regarding plasma....well i like it so much but i recently known that is not a multithreaded app(am i wrong?),if i'm not i guess it really needs to be ported/rewritten to support multithread.


By andrea at Tue, 2010/12/21 - 10:57am

...but you really must learn to use punctuation properly, otherwise it is extremely difficult to read what you write.

I can't agree with you about kmail. It is fraught with issues. (See below.)


By Dulwithe Darkingham at Wed, 2010/12/22 - 7:09pm

I agree with the Metadata problem.

They must be embedded in the file, storing in an registry is an error some other Software editor did before kde.
I will never spend time adding some meta information on a file if i know they will be lost on a simple 'mv' operation.

I have a NAS server, used in a heterogenous OS environment, this is going to be very common, and a registry on a local machine is NOT the solution.

best regards


By lbayle at Tue, 2010/12/21 - 2:09pm

e.g.: all the buzz about Akonadi is great, but I never got a good reply to my question what to do with Akonadi in an environment where $home is distributed over NFS?!
Akonadi runs a mysql instance locally and the database is also under $home. But having a mysql db on an NFS share is not exactly a good idea. This is not some rare and unlikely use-case...c'mon... that used to be one of the advantages of Linux/KDE!... now it's going in the opposite direction. Not everybody is a private user with only one single PC at home.


By thomas at Tue, 2010/12/21 - 3:04pm

Akonadi also supports remote db, where the MySQL instance resides on a server on the network. it could be on the same server as your NFS shares, for instance. then the MySQL I/O happens local to the data.


By Aaron Seigo at Tue, 2010/12/21 - 8:08pm

I didn't know that! Than it's even better than before!
That should also make backups a lot easier! Great! Thank you!


By thomas at Wed, 2010/12/22 - 9:32am

that's a good point, but don't solve the problem if you use environments where Akonadi is not deployed.
I realy think metadata should be embeded in the files. (on this point of view mp3 is great, don't you think so ?)


By lbayle at Wed, 2010/12/22 - 5:45pm

i addressed the metadata locality issue in another comment and i do think it is worth effort being put into it, but this was about the issue of MySQL doing I/O on the local machine on a network volume and how to avoid that. since akonadi itself is "just" a centralized data cache and doesn't actually create metadata, there isn't much to say here bout metadata locality :)


By Aaron Seigo at Wed, 2010/12/22 - 6:27pm

I think we should make a list with all this crucial stuff we need to fix before starting to do new stuff for 2016, on my list:
-Make dolphin open instant and respond better.
-Make printing not suck, reviving kprint and completing the printer administration sysconfig module.
-Make overall desktop performance/responsiveness better, I think qt is the principal culprit here but kde also has a lot of work to do.
-Improve metadata handling, this one seems hard, I have no idea what would be a better solution
-Improve rekonq experience, maybe vp8 should be used, also scrolling performance is really important for me, apart from that it's actually getting really good.
-clean sharp borders in apps, specially ui. We should start looking more to the usability


By damian at Tue, 2010/12/21 - 3:28pm

add to that:

-a messaging application supporting _reliable_ audio and video chat (it could be based on PSI). Let's ditch all other protocols and offer complete, solid and consistent support for the open one: Jabber
-kio-slaves need some serious work. they should be as fast as the command line
-separate nepomuk and akonadi from the core, make them optional packages. for many people they're nothing more than an annoyance.
-dolphin needs lots of usability improvements and some degree of awareness to reduce its overhead, specially when you're using it over kio.
-krunner needs some love (this is an easy one)
-an image-viewer that opens instantly. I mean _instantly_. It doesn't need to have many functions, just display the image, zoom it and navigate a directory.
-a decent, complete, modern and fast pdf-reader. okular is currently in a deadlock and knowing Pino, it will stay that way.
-kile would be wonderful if it had a nice spellchecker.
-kdetv was great and I miss it
-discontinue koffice, it's a waste of effort
-a replacement for "ark" is desperately needed
-remove "activities" from core, offer it as a plugin, like superkaramba used to be


By trixl at Tue, 2010/12/21 - 6:04pm

> -a decent, complete, modern and fast pdf-reader.

Could you tell in specific those? Otherwise, as generic they are they mean basically nothing.

> okular is currently in a deadlock and knowing Pino, it will stay that way.

Thanks for the gratuitous attack, Mr Anonymous. Who are you to declare you know me?


By pinotree at Tue, 2010/12/21 - 7:10pm

i'll just touch on the items i have anything at all to do with; for many of the rest of the items on your personal wish list, i suggest participating so that what you want happens.

"-separate nepomuk and akonadi from the core, make them optional packages. for many people they're nothing more than an annoyance."

and for the rest of humanity, the only way to get their benefits is to make them available. if we end up satisfying the needs of the rest of humanity at the expense of a few people such as yourself, that's fine. our job is not to create the ultimate environment for absolutely everyone, but to create a coherent and consistent experience that enables a given set of uses.

"-krunner needs some love (this is an easy one)"

it gets love with each release. i wish there were more people who worked on it, but we have only so many resources. it improves with each release, and i hear more and more often how people have come to rely on its functionality which is a good sign. in any case, the answer to this one is "more people and/or more time" :)

"-discontinue koffice, it's a waste of effort"

you are evidently operating under the assumptions that by doing so those people would start working on some other things in KDE. (and, i assume, you'd hope for them to work on your personal pet issues, no less. :)

well, that's simply not true. if koffice were discontinued, the majority of those people would simply stop contributing altogether. this is an implicit aspect of participation based creation.

furthermore, KDE has received a lot of improvements, particularly in kdelibs, due to koffice over the years.

"-remove "activities" from core, offer it as a plugin, like superkaramba used to be"

a) in Portugal they have rolled out 100s of 1000s of laptops with KDE on them. they used Superkaramba previously for their needs, and with Plasma they moved to Plasmoids. they had significant issues with Superkaramba, issues that spring directly from it being an add-on rather than an integrated feature. these issues, to the last one, disappeared for them with Plasma. i'm sure they, and others who are similarly benefiting, would be immensely pleased if we took your advice and screwed them over.

b) every single component on your panel and desktop are driven by this set of features. just as kicker in KDE3 had applets, Plasma has Plasmoids. if we removed those features, we'd simply have to replace them with yet another plugin system, meaning we'd be right back to where we started. unless, of course, you don't want to configure your panel or you feel that you don't need a launcher, system tray, task bar, etc. at all.

:)


By Aaron Seigo at Tue, 2010/12/21 - 9:55pm

>b) every single component on your panel and desktop are driven by this set of features. just as kicker in KDE3 had applets, Plasma has Plasmoids. if we removed those features, we'd simply have to replace them with yet another plugin system, meaning we'd be right back to where we started. unless, of course, you don't want to configure your panel or you feel that you don't need a launcher, system tray, task bar, etc. at all.

I'm not sure but I think he meant activities like in activity manager, not plasma. Activities are largely not understood, they have a good purpose but it is hard to understand, maybe making the ui more intuitive would help.
I've had an idea about it(I'll do a mock up some day)
This is a bit offtopic but on the comments of this post there is a lot of info, My idea is in the comments with the same user.
http://chani.wordpress.com/2010/12/18/activities-4-6-screencast/


By damian at Tue, 2010/12/21 - 10:11pm

the comparison with superkaramba, as well as the use of the term "activity" previously in plasma-desktop where we now use "widget layout" makes me think they meant widgets :)

still, activities themselves are built on that same substrate and removing them would do absolutely nothing of benefit.

we're working on various aspects of the UI, yes. will likely replace what is there now with something QML based in 4.7 (Marco already has a basic mockup in QML done). more hands helping out are always welcome :)


By Aaron Seigo at Wed, 2010/12/22 - 6:25pm

>"support for the open one: Jabber"
I think kde should support every protocol, and do it right, it would be one step at a time but sharing code with other projects it should be possible, telepathy should help
>"a decent, complete, modern and fast pdf-reader. okular is currently in a deadlock and knowing Pino, it will stay that way."
I don't know why everybody is saying so many bad things about okular, It has served me really well it's a fantastic app, it could use some performance improvements, specially for note-taking. And maybe a few killer features, but (at least with pdf) it's a great app anyway. If you don't like how it renders something, libraries should be the culprits, format rendering is usually done by libraries, so libraries are the one who should get better(ghostscript for example)
>"-discontinue koffice, it's a waste of effort"
I really don't think so, krita and karbon are great apps for example, and the others are steadly moving forward, he new calligra renaming gave them some momentum too.
>"-a replacement for "ark" is desperately needed"
Could you explain why? it never gave me a problem.
>"-remove "activities" from core, offer it as a plugin, like superkaramba used to be"
It would be impossible to do that AFAIK, because they are too buried, what could be done if normal user is overwhelmed by another layer, is disabling/enabling them in sysconfig.
>"separate nepomuk and akonadi from the core"
Are you talking of space, or does it just feel dirty? (it happens to me a lot), they don't take up that much space and you can disable them so if it's not performance (if they are disabled they don't consume anything) is it space? (if kde is used in mobile devices where space is limited,etc this might be important) is it dirtiness?
I agree, tough, with all other things you said, and I think the instant image viewer should be the same gwenview, maybe optimizing to open the image fast, and display it while loading the rest of the interface and the tools.


By damian at Tue, 2010/12/21 - 10:04pm

>Could you explain why? it never gave me a problem.

Hard to explain, but ark always feels really clunky to me. Part of it's the way the UI freezes on load (probably inevitable due to the tar format, but it makes it feel unresponsive). Other than that... everything just seems to be too many clicks. What I usually want is to have a quick look at the contents of an archive, and then extract it to the same directory the archive is in, so perhaps this particular workflow could be optimized.

>Are you talking of space, or does it just feel dirty? (it happens to me a lot), they don't take up that much space and you can disable them so if it's not performance (if they are disabled they don't consume anything) is it space? (if kde is used in mobile devices where space is limited,etc this might be important) is it dirtiness?

My biggest problem with akonadi is when it breaks (which is frequently). The most hilarious problem this causes: if you start up kmail, you get a warning box about how some part has broken, and when you click close it closes kmail. But, if you just drag this dialog out of the way, you can use kmail as normal - it's perfectly functional. Which leaves me wondering what the hell this stuff actually does, other than get in the way of actually using my programs.

As for image viewer, kuickshow gives me everything I need and is instant as required. So I don't think this is something the kde project needs to work on.


By lmm at Wed, 2010/12/22 - 3:01am

kmail 1 uses akonadi to access adress book, so if you had an error in akonadi, some of the contacts may not be accessible.
If you don't like akonadi you can use a mail client that doesn't use it and totally disable nepomuk/akonadi, removing it physically from the disk won't help with problems AFAIK, it woudl only free a tiny bit of space.


By damian at Wed, 2010/12/22 - 2:39pm

>an image-viewer that opens instantly. I mean _instantly_. It doesn't need to have many functions, just display the image, zoom it and navigate a directory.

Try kuickshow


By lmm at Wed, 2010/12/22 - 3:02am

it isn't available on the opensuse-repository. I wonder why...


By trixl at Thu, 2010/12/23 - 2:20pm

I also would like to use kuickshow on Ubuntu (10.04 LTS).
(In my opinion this was the best picture viewer I've ever used. Certainly the user interface does not conform to any standard. But there is usually no need to interact with the program in another way than opening a file. The program is doing "the right thing" without need of additional interaction.)
But it is not listed in the "Software Center".
Does anybody know where to tell "the Ubuntu-people" to change this?
Any hints wellcome! Many thanks in advance.


By JoeHurke at Sat, 2010/12/25 - 2:01am

I think KOffice should stay - it is becoming more important on mobile platforms because it is smaller and faster than OpenOffice / LibreOffice.

What I would like to see is the Konqueror web engine becoming more compatible - it is fairly good, but not perfect. I love the "clone tab" and "detach tab" features, but some websites don't work, and then I have to switch to firefox.

I do not know if this should be KDE, or rather an X-windows improvement, but the ability to move the display of running applications from one machine to another (for example to a tablet PC or a colleague's PC).

Kioslaves needs more logical names and maybe a dropdown list. Teaching new linux users to remember fish:// or smb:// is not easy. And it is less reliable than commandline.

Dolphin is almost on par with Konqueror as a file manager, but some settings such as "Show delete button" is still not available from dolphin, only from konqueror.

Personally I think YaST from SuSE should be integrated into KDE. I know some people don't like it, but I don't like a different configuration tool for everything.

The new functionality where you can group windows, should also group the taskbar icons, and not showing taskbar items from other desktops should be the default to reduce clutter. If a taskbar shows windows from all desktops, the concept of multiple desktops is pointless.


By JHB at Wed, 2010/12/22 - 9:34am

That should be added to the list too, about yast, maybe yast isn't the best for every distro, I think configuration managers need to integrate into kde sysconfig by making kcm, and if it's a general one (not distribution dependant) then promote it to base kde.


By damian at Wed, 2010/12/22 - 2:42pm

Yeah, printing is a big issue.

It takes too many clicks to go through and set printer settings. And why are there 2 different page size settings settings (one is called "media")?

The kde4 print dialgue is bizarre and other than the "print assistant" available in gwenview and digikam (I think it is also in digikam), I find it extremely problematic to print photos well.

Also, things like kmail messages getting truncated and other info getting truncated (ie: cut-off). Leaves a user frustrated.

Again, as per my point below, kde team needs to decide "are we gonna make this a hobbiest/enthusiast/developer project with loose ends all over the place, or are we gonna narrow our focus and get down to business creating a DE and software ecosystem that actually works well for the user??" This is the decision the kde team must make.


By Dulwithe Darkingham at Wed, 2010/12/22 - 7:19pm

I agree that the kdepim suite is unfortunately in a bad shape at the moment - hopefully that will change with kmail2 whenever it will be released. As far as metadata are concerned, a major annoyance for me is that nepomuk is still going wild from time to time and consumes way too much cpu power. But slowly some of its benefits are showing up, like the improved search features in kde 4.6's dolphin. Okular works well as a reader in my opinion, but extra functionality like embedded annotations/marks would be very welcome.

Last but not least, one should also not forget about those impressive applications that work wonderfully well, like for example amarok and digikam!

So although I, as a non-developer, am confused that progress is so slow, I have the feeling that the whole kde suite is constantly improving and maturing, and count me as a happy user :-)


By stpsi at Tue, 2010/12/21 - 4:22pm

"Do we really need yet another layer on top of that?"

no, and it isn't another layer on top of that. activities are an entirely orthogonal concept to each and every one of the examples you gave.

"I know that the idea behind it is to "keep the state" but almost everyone use the same programs everyday so it doesn't really matter."

some people do, yes. many people do not. this is a result of people carrying their computers around with them physically as well as the growing number of topical contexts people use those computers for.

"And how about the obsession with metadata for everything?"

since it lets us work with our information in not just better but also new ways, it's great. as a result, people use it all over the place. tagging people in photos or tagging photos with the location (automatically or manually) is already a well established pattern, to pick just one of hundreds.

"You cannot copy a file with its metadata using the command line, through FTP or a web server, or any other non-kde way. This is a trap"

it's not a trap, it's something that can be improved on. the answer to "it's not as good as it could be" is not "don't bother trying" but "improve it". i mean, wasn't that the mantra of your message in the first place?

thing is, without things that use metadata, metadata improvements lower in the stack are so totally unlikely to happen that we can just assume they won't ever. so we work on the user benefit parts first, discover what needs improving, and move on to those things.

"The things that are really useful and matter, seem to be the ones receiving the less attention of all."

some things definitely could use more attention. printing and kopete are two you mention, and i agree. that attention doesn't come from limiting other avenues of progress, however. that just means nothing improves anywhere. many of KDE's contributors come in and start improving things because they are first attracted by other useful features. and those working on those other useful features are ensuring that those parts of KDE, which are also important to many, stay up to scratch. your entire analysis of the situation is wildly oversimplified.

"go and hang Aaron"

the day you walk up to me and make that threat to my face is the day you earn the right to post that kind of thing behind the comfortable distance and anonymity of your computer. until then, that kind of personal threat is juvenile and cowardly. i doubt you'd have the temerity to say such a thing to me in person, however, and that perhaps would give you something to think about.

"then care about the _present_ of KDE."

many of us already are caring for the present of KDE. that's why it improves from release to release, and not just in terms of new features and capabilities but improvements on what already exists. if you are wondering where those might be, follow KDE blogs for a while or, if especially motivated, the commit mailing list.

this doesn't absolve us of the responsibility to also care about what the long term future of and plans for KDE are. we have to do both: improve the now, set direction for the future.

for a very nice example of why this is so, here's a fun experiment (though make sure you have someone with you to prevent disaster): go to a store (preferably in a neighbourhood you are not familiar with) to purchase a food item, say a carton of milk. before you start, put plugs in your ears. then as you travel, only look directly at your feet and speak to no one as you go. never look up, never look to the side. it's pretty hard to do, and chances of succeeding in getting the item are pretty low. and worse than failing to get the milk, if you wander around like that long enough you will almost certainly eventually end up in an accident, potentially a life threatening one.

we have to be able to look forward, just as we need to also know the ground we are standing on right now.


By Aaron Seigo at Tue, 2010/12/21 - 9:43pm

>the day you walk up to me and make that [...]

I was speaking figuratively, in case you didn't realize. I oppose the death penalty, even in your case Aaron ;)

Btw. I'm not opposed to the use of metadata but I think that it should be stored and handled by either the file system or the file format and _not_ by an external application storing it separately.


By trixl at Tue, 2010/12/21 - 10:55pm

This is not something to joke about.


By Lydia Pintscher at Wed, 2010/12/22 - 4:17pm

"I was speaking figuratively"

figuratively or not, it isn't a rational argument it's lashing out at someone personally. it's way out of line, and i'd encourage you to address facts and issues rather than people you don't even know.

and since it evidently isn't apparent to you: comments like yours cost free software projects contributors all the time. most people, reading the kind of crap you wrote, find it disheartening and a good reason to stop participating. i can't blame them. if you've read planetkde in the last year you'll have seen at least one person announce exactly that and then leave. they are not alone. it is one of the main causes of attrition to the contributor community, in fact. and if you think your comment hasn't had an effect, i've received multiple private responses, all disheartened, after having read what you wrote.

to put a fine point on it: you are helping kill free software by destroying that which helps people want to be a part of this in the first place: the community and camaraderie. on behalf of every one of my friends and associates who have walked away because of people like you, i find your comment and what it says about you to be despicable.

if you care about the future of free software, you need to significantly rethink how you approach these things. and i'm not asking you to consider becoming a sycophant, i'm asking you to consider being civil and kind to others.

"I think that it should be stored and handled by either the file system or the file format and _not_ by an external application storing it separately."

that is a highly unrealistic expectation right now. it's something to shoot for, but it's not going to happen right now.

it also ignores the facts that: this information needs to be indexed for quick search (can in theory be done in the file system; experience shows that has performance and stability downsides, however); not all metadata reflects the contents of a file (though the metadata itself can be represented as a file); the relationships between pieces of metadata is as important/useful as the metadata itself, meaning that storing it in the file alone is not enough and even in the file system is likely to be too constraining.

now ... what we have done is create one central place for metadata (nepomuk) and a common API to use to access and store it. this is a huge step forward from the case just a few years ago where this metadata was still being created and stored, but by each individual application in their own place. this allows us to repurpose the storage system in future much more easily as well as opens the doors for the relationship-between-metadata graphs to be built and subsequently queried.

it's quite evident that the people working on this actually do know what they are doing. i'd really encourage you to do some deep research on the topic, as that's probably the best cure for misunderstanding.


By Aaron Seigo at Wed, 2010/12/22 - 6:50pm

I'm a monster Aaron. I'd write an apology, but I've been running my laptop on battery for like 5 minutes and kde-4.6ß2 already ate almost all the battery power. Perhaps tomorrow. Tonight I'll cry.


By trixl at Wed, 2010/12/22 - 11:12pm

You spent your 5 minutes of battery power writing that? What were you thinking?


By tangent at Thu, 2010/12/23 - 4:22am

That last sentence is _way_ out of line and would there not be more useful answers to your post I would delete it.
This is not ok here!


By Lydia Pintscher at Wed, 2010/12/22 - 4:11pm

Well.. it's actually only rumors...
Nokia is going to switch to Microsoft on the mobile front. No Meego /Maemo anymore... Qt gets kinda obsolete..
Think they just failed. Never put enough development into Maemo nor Meego.. instead continuing on Symbian. Big mistake.. Now getting bought out by Microsoft and Qt/Trolltech gets obsolete.. nice shift...


By thomas at Tue, 2010/12/21 - 12:47pm

had the same thoughts when reading the news... lets hope thats no the end for qt. i'm not saying they will kill it quickly, but perhaps let it die slowly with less and less development


By Beat Wolf at Tue, 2010/12/21 - 12:51pm

the former Trolltech company is now stuck in the middle... no real future left.
There's no need for Nokia to stay with the frameworks they have been developing in the past. Microsoft delivers a complete and consistent stack of development tools for its platform.
The new Nokia CEO seems to be a "smart" guy.
Nokia has two chances: either get Meego up to speed (the best thing they could do from my POV) or start from zero with Windows.


By thomas at Tue, 2010/12/21 - 1:12pm

QT could be forked before it gets native SilverLight support.


By trixl at Tue, 2010/12/21 - 1:04pm

...they are used not only by Nokia for develop Meego or Symbian or whatever, are used by thousand of applications, proprietary or not, Adobe, Autodesk, Google, Samsung, Panasonic, Philips, Oracle, are just some of the company that uses Qt for develop applications. Qt are a good business for Nokia.


By Zell_89 at Tue, 2010/12/21 - 2:04pm

was to push Maemo/Meego... somehow it did not work out. Probably the development was not fast enough. Dunno the reasons why Nokia did not try to switch to Maemo/Meego. Can be because Qt was just not up to the expectations, so they _had_ to stay with Symbian for too long and afterwards it was too late so they _had_ to take Windows...


By thomas at Tue, 2010/12/21 - 2:17pm

I doubt that Nokia intends to give up their software revenue. Remember at present Symbian is the most popular smartphone OS world wide because Nokia is the biggest mobile phone manufacturer. Whatever they adopt wins so MS really wants to court them for their smartphone OS rather than stay a minor major player against Apple, Google and RIM in the great smartphone wars of the twentyteens.

I don't think their QT/Meego plans are in danger but I do wonder how such major players as Nokia and Intel have taken so long with their OS as to have dropped the ball. This negotiation is probably a result of Nokia waking up to that fact.

Nokia is the biggest because of quality hardware and I don't see any competitors there so whatever they end up with people with buy but a stopgap solution may be necissary in the short term so customers aren't swayed by other companies press.


By Shaun Hunter at Tue, 2010/12/21 - 4:28pm

Thats what this article asked for not a KDE is dead gripefest. Over 5 years major underlaying changes can be made so a roadmap should be put down. The GNOME policy of nobody moves nobody gets hurt isn't very good imho. KDE4 will get incrementally better so don't sweat the small stuff.


By Shaun Hunter at Tue, 2010/12/21 - 4:31pm

I hope KDE 4 will be someday as usable and stable as KDE 3.5 was.
Looks like it will take at least till 2014 to get this.
But perhaps it might take till 2016.
So in my opinion this would be a great goal.


By JoeHurke at Wed, 2010/12/22 - 2:11am

I still have a couple of computers running KDE 3.5. So I get to compare them as they are, rather than relying on some rose-colored memory of what I think KDE 3.5 was like. And I have to say, as KDE4 has advanced the annoyances of KDE3 have become more apparent. For most things, I'm a much bigger fan of KDE 4.5. Not to say that there is nothing that 3.5 didn't do better, but on average it's been a huge step forward.

Honestly people, discussion is like painting. When you paint a picture, it doesn't make sense to use brushes that are bigger than the canvas itself. A fine-scale brush lets you be honest to the detail that is present in the scene you are trying to paint. No need to wipe out the Van Gogh when all you need to do is touch it up.


By tangent at Wed, 2010/12/22 - 11:01am

I've used KDE 3.5 (on openSuSE 11.1) as "productive" environment at my home office till May 2010.
So I've no "rose-colored memory" about KDE 3.5.
Then I had to switch to Gnome because Kaffeine crashed every few minutes on KDE4 (KUbuntu 10.4 LTS).
I had to switch from Konquerror to FireFox because mutch of my bookmarks are unreachable with Konquerror.
I had to switch from KOffice to OpenOffice because KOffice crashed every time I tried to open any of my KOffice-SpreadSheets.
I'm still using KMail, though it duplicates Mails every time I download Mails https://bugs.kde.org/show_bug.cgi?id=259178 and filtering based on AdressBook is not working https://bugs.kde.org/show_bug.cgi?id=259181 .
Perhaps I will have to switch to Thunderbird because of these problems.
I'm still using Kaffeine though slow motion and fast forward have gone.
All these problems came with KDE4. All this was working with KDE3.
I'm not talking about the unreadable system messages in KDE4...
I'm just trying to keep a usable and stable home office environment.

What does KDE4 need beyond "working applications"?

* reorganization of settings (so that I don't anymore have to search everywhere to find anything)
* application start panel that can compete with Gnome3 (more overview at a glance than now)
* at a glance readable system-messages (no resize to read file names, no clicks to open progress bar of file copy, no clicks to confirm completed file copy that I have already viewed)


By JoeHurke at Fri, 2010/12/24 - 1:39am

Don't worry. Everything is part of God's plan.


By trixl at Fri, 2010/12/24 - 8:41pm

While I like polish as much as the next person, I can't agree with the tone of many of the comments here. I am really looking forward to upcoming versions of activities, nepomuk search, etc. People who think these things need to be abandoned and ripped out have not been noticing how steadily KDE has been improving since 4.0. Heck, it's the one component of my computing experience that I can guarantee will get better each time I upgrade. Giving up now, when we are so close, is juvenile.

Since all things take time to mature, now is the perfect time to be thinking about more long-term innovation. We're only now harvesting the benefits of the vision from five years ago. While feasting, we also have to plant the next generation of crops.

How about:
1. Dictation? Is there a good open-source speech-to-text engine?
2. I keep wanting to find ways to "push" a window over onto someone else's desktop: "hey, check this out." Maybe this already exists and I don't know about it.
3. How will 3D displays change things?


By tangent at Wed, 2010/12/22 - 2:48am

Pages