Linked by Thom Holwerda on Wed 3rd Jan 2007 17:54 UTC, submitted by ronaldst
KDE "Since KDE 4 development is in full swing with plans for a KDE 4.0 release sometime later this year, I thought I'd put together a weekly piece entitled The Road to KDE 4. The idea is to have a short overview of one or two of the features that show progress in KDE 4. For my first issue, the goal is to show off some of the great SVG work that has taken place so far."
Order by: Score:
nice
by backdoc on Wed 3rd Jan 2007 18:10 UTC
backdoc
Member since:
2006-01-14

Since switching to Ubuntu, I've gone from being a KDE fan-boy to predominantly using Gnome and XFCE4. I like Gnome and XFCE4 but I miss KDE. KDE 3.x has just lost its appeal to me. I'm excited about KDE4, though. The screenshots look great. I'll be looking forward to KDE4.

KDE gives me the control over my DE that I enjoy. Gnome gives me much nicer defaults, I guess.

Reply Score: 5

RE: nice
by bedo on Wed 3rd Jan 2007 23:42 UTC in reply to "nice"
bedo Member since:
2006-01-03

switch to Kubuntu

Reply Score: 2

RE[2]: nice
by backdoc on Thu 4th Jan 2007 00:38 UTC in reply to "RE: nice"
backdoc Member since:
2006-01-14

I've used Kubuntu. It's OK. But, I just feel more "comfortable" with Gnome these days.

Reply Score: 2

elsewhere
Member since:
2005-07-13

...at least IMHO.

Before people judge one way or the other, remember that these are still developmental and in no way represents the final look 'n feel KDE 4 will have, they're more to show off just how the new graphical elements can be used and applied.

There is some cool work being done on the interface level, from what I gather following the dev blogs this is going to give excellent flexibility for designing and implementing application/GUI layouts. The use of CSS for styling etc. should also allow a better seperation between the application code and the interface design.

Kmahjongg was due for an update, no doubt about it and the look is certainly improved, but I think to look at something like ksystemguard demonstrates how signficantly the presentation of the app can be improved without necessarily adding complexity.

I've always thought screenshots were a bit overrated, but in this case, particularly in comparison to the 3.5.x equivalents, a picture is worth a thousand words.

Reply Score: 5

DigitalAxis Member since:
2005-08-28

Right now I'm hoping that if KDE4 goes entirely SVG, we'll get to see things like resolution-independent displays.

Reply Score: 4

GreatBunzinni Member since:
2005-10-31

Indeed it has the potential to be a great thing. As we can see, things will get increasingly easier on the eyes and certain aspects of writing code will get somewhat easier. Those are indeed good things.

Nonetheless, all the SVG rendering comes at a price and that aspect worries me. Althought KDE3 is very responsive and lightweight enough, will all that processing needed to render SVG graphics slow KDE down? It would be a shame if KDE's bloated reputation ceased to be a FUD myth and started to move to the realm of facts and reality.

Reply Score: 3

MightyPenguin Member since:
2005-11-18

SVG Rendering should only have a startup cost, after that they should be cached and only updated if rescaled. So slightly more memory consumption but basically the same after-startup performance.

Reply Score: 1

Gooberslot Member since:
2006-08-02

That might be true for some things but what about things like the system monitor that are constantly updating.

Reply Score: 1

superstoned Member since:
2005-07-07

well, that's mostly anti-aliasing and such stuff, not that heavy. and besides, slowly more and more effects will be accelerated by OpenGL, so effectively it might even become faster... Qt 4.3 is supposed to use OpenGL for much more drawing operations, speeding up for example SVG.

Reply Score: 3

JohnFlux Member since:
2007-01-04

I render the SVG to an image. When the graph is updated, it simply bitblits the image. It's very fast and probably comparable to just drawing a black background

Reply Score: 4

Great idea! Keep 'em coming!
by Fusion on Wed 3rd Jan 2007 18:18 UTC
Fusion
Member since:
2005-07-18

This article series is a great idea. Digging through KDE developer blogs and SVN changelogs just to 'feel out' the progress on KDE4 is a pain, often yielding more questions than anything else. Having a regular, predictable, and quantifiable presentation of KDE4 progress certainly fills a gap on this site.

OSNews.com should consider creating a news specialist position for each of the major desktops. It'd be nice having an 'expert' (or at least someone well-versed) make regular news contributions on desktop progress. That'd free developers up from being squeezed into playing public relations, so they can get back to doing what they do best---coding! =)

Reply Score: 5

RE: Great idea! Keep 'em coming!
by sbergman27 on Wed 3rd Jan 2007 19:02 UTC in reply to "Great idea! Keep 'em coming!"
sbergman27 Member since:
2005-07-24

"""This article series is a great idea."""

Indeed.

Interesting. Informative. Positive.

And oh so much more productive than "armchair editorials", which breed so much useless "negative energy" for lack of a better term.

Edited 2007-01-03 19:02

Reply Score: 5

Looks great
by Shaman on Wed 3rd Jan 2007 18:19 UTC
Shaman
Member since:
2005-11-15

When KDE4 shows up, it is going to be miles ahead of just about any other environment out there. Not necessarily in feature set per se but in overall synergy. Very nice.

Reply Score: 5

RE: Looks great
by nutshell42 on Thu 4th Jan 2007 16:15 UTC in reply to "Looks great"
nutshell42 Member since:
2006-01-12

When KDE4 shows up, it is going to be miles ahead of just about any other environment out there. Not necessarily in feature set per se but in overall synergy. Very nice.

Today on Osnews: "When KDE4 is released its buzzwordiness will empower a Gnu generation!!"

Coming soon: Osnews's article "Minor roadblock for KDE4" with the hit reply "KDE 4 is doomed! Doomed!"

A bit less hype/fud please, pretty please.

Reply Score: 2

All I can say is...
by djst on Wed 3rd Jan 2007 18:27 UTC
djst
Member since:
2005-08-07

...Wow!! Looks like I'll be using KDE when 4.0 comes out, something I never thought would happen.

It's a shame though that Firefox, Thunderbird and OpenOffice.org (all apps I use extensively) will probably look extremely out of place in KDE4.

Reply Score: 4

RE: All I can say is...
by Julikaefer on Wed 3rd Jan 2007 20:15 UTC in reply to "All I can say is..."
Julikaefer Member since:
2006-08-14

Who needs Firefox, Thunderbird and OpenOffice.org in KDE4 when there is Konqueror, KMail and KOffice? ;)

Reply Score: 5

RE[2]: All I can say is...
by GreatBunzinni on Wed 3rd Jan 2007 20:26 UTC in reply to "RE: All I can say is..."
GreatBunzinni Member since:
2005-10-31

To be honest, KOffice, in it's current 1.* stage, sucks too much. It is unusable. For example, KWrite isn't even capable of handling tables correctly, Karbon crashes when copying drawing objects, KSpread crashed (until the last release) when changing cell values, etc etc etc... Possibly KOffice 2 will solve all of that but as it is very far away and it has yet to pass the dreaded 2.0 release, I believe it will be a long while until KOffice is remotely viable.

On the other hand KMail and konqueror are very usable. Yet, their mozilla counterparts are still miles ahead of them at least in terms of usability. For example, knode's post window crashes when changing dictionaries and, even after a couple of years, it doesn't support marking messages for download.

So, to sum things up, KDE's application package is slowly improving but it still leaves much to be desired when compared with the leading applications. Some applications may be usable but the overall experience still sucks a bit.


/exclusive KDE user since 98

Reply Score: 4

RE[3]: All I can say is...
by superstoned on Wed 3rd Jan 2007 22:25 UTC in reply to "RE[2]: All I can say is..."
superstoned Member since:
2005-07-07

sure, but many of these problems will finally be fixed by the Koffice 2 release - thanx to the better infrastructure in Qt 4.x and the work they are doing on their libraries. ok, it'll still be behind in some stuff, but i'm sure the basics will work much better.

Reply Score: 5

RE[4]: All I can say is...
by Hiev on Thu 4th Jan 2007 00:09 UTC in reply to "RE[3]: All I can say is..."
Hiev Member since:
2005-09-27

I don't think crash problems, bad tables support, etc, are problem of the framework but the code of KOffice itself, that's like saying KOffice actual problems are for the problems of Qt3 and that's not the case.

Reply Score: 2

RE[5]: All I can say is...
by superstoned on Thu 4th Jan 2007 07:41 UTC in reply to "RE[4]: All I can say is..."
superstoned Member since:
2005-07-07

the crashes probably not, they are mostly mistakes and need time to be fixed, but the bad table support is surely hard to fix without Qt4, same with the text layout problems. and for example grammar checking will be in KDE4 so Koffice will have that, and they don't have to spend (much) time on it.

i've been reading the mailinglist, and sometimes compiling the KDE4 Koffice SVN and it again and again you see them implement things they wanted for a long time but where hard to do. just because they almost come for free in the new framework they're building...

Reply Score: 2

RE[3]: All I can say is...
by zombie process on Fri 5th Jan 2007 01:00 UTC in reply to "RE[2]: All I can say is..."
zombie process Member since:
2005-07-08

I'd have to agree about koffice, but kmail and konqueror are awesome, IMO. Notice I said kmail, not kontact which needs some attention - loading stand-alone apps in a nested window != a unified app. I'm really hoping that some energy is spent here - I'd really love a better thought out addy-book and cal that I could use easily within my email program.

Reply Score: 1

RE[2]: All I can say is...
by djst on Thu 4th Jan 2007 11:48 UTC in reply to "RE: All I can say is..."
djst Member since:
2005-08-07

Who needs Firefox, Thunderbird and OpenOffice.org in KDE4 when there is Konqueror, KMail and KOffice? ;)

Because I dual-boot, I can share my bookmarks, form login, passwords, etc. between Linux and Windows XP using Firefox.

Similarly, I can share my mail, contacts, etc using Thunderbird.

I can't comment on KOffice as I haven't used it.

Reply Score: 3

RE[3]: All I can say is...
by zsitvaij on Thu 4th Jan 2007 11:53 UTC in reply to "RE[2]: All I can say is..."
zsitvaij Member since:
2006-06-14

Given that KDE4 will be fully cross-platform from the start, you can do all of that with Konqueror, Kontact/KMail, KOffice, Kopete, Digikam, Amarok, etc.

Reply Score: 1

RE[4]: All I can say is...
by pospiech on Thu 4th Jan 2007 12:03 UTC in reply to "RE[3]: All I can say is..."
pospiech Member since:
2007-01-04

KDE will have cross-platform ability as a platform, but the applications will only be crossplatform as soon as they have been ported - as far as I understand it.
So KDE 4.0 might have very little crossplatform applications, but later releases should show all the magic KDE 4 can do.

Reply Score: 1

RE[5]: All I can say is...
by hal2k1 on Thu 4th Jan 2007 13:02 UTC in reply to "RE[4]: All I can say is..."
hal2k1 Member since:
2005-11-11

//KDE will have cross-platform ability as a platform, but the applications will only be crossplatform as soon as they have been ported - as far as I understand it.//

As I understand it, "porting" a KDE 4 application will be no more difficult than "recompile". It may not even need a "recompile", it may be possible to just "re-link".

KDE4 apps will have dependencies on KDElibs and on QT4. As I understand it, both of those are cross-platform. There is already a QT4 for Windows, AFAIK.

http://www.google.com.au/search?q=%22QT4+for+Windows%22&hl=...

It would appear that there is even a GPL version of QT4 for Windows.

Reply Score: 2

RE[6]: All I can say is...
by elsewhere on Thu 4th Jan 2007 14:30 UTC in reply to "RE[5]: All I can say is..."
elsewhere Member since:
2005-07-13

KDE4 apps will have dependencies on KDElibs and on QT4. As I understand it, both of those are cross-platform. There is already a QT4 for Windows, AFAIK.

Actually, Qt is more widely used on Windows than it is in Linux/X11. The big advancement for cross-platform capability was Trolltech's decision to GPL Qt 4.x for Windows/OSX, previously they were commercially licensed only.

This is part of the beauty of KDE's new design, not just as a desktop environment, but an application environment. Some of the framework designs, like Solid, are designed specifically to "untether" KDE from being tied to a specific platform when it comes to things like hardware interfaces. Qt on Win/OSX permits apps to run with more of a native appearance, and not requiring a workaround like cygwin or fink.

So there will be porting involved for things like Solid to interact with the hardware interfaces on Windows (or OSX), and something like Phonon which will probably simply use native sound systems as backends (again, transparently to the KDE-based apps). But with that, you're correct that there should theoretically be little "porting" involved for KDE apps once the core libraries are available natively on other platforms.

Right now, the discussion I've seen is around Konqueror, KOffice and Amarok being ported to non-*nix platforms, in fact one of the devs is maintaining development builds of KDE4 apps for OSX already. An app like Kontact should be even easier to port since there's little platform-dependent "complexity" involved.

I guess a decision will have to be made as to whether Win/OSX users will download and install a KDE4 framework in order to use these apps, or the framework will be statically linked to the individual apps to provide standalone functionality.

The goal ultimately is to emphasize the apps and diminish the relevance of the host platform, some people think that this can undermine linux by providing high-quality free apps for Windows (and OSX), but I take the viewpoint that it strengthens the relevance of desktop *nix by reducing dependencies on closed platforms.

Time will tell...

Reply Score: 3

RE[2]: All I can say is...
by panzi on Wed 3rd Jan 2007 21:17 UTC in reply to "All I can say is..."
panzi Member since:
2006-01-22

They don't look that much out of place. With the Qt-gtk-engine Gtk-apps (as Firefox and Thunderbird) look almost like Qt apps. And there is also a Qt/KDE engine for OpenOffice. Just try the newest Kubuntu and see what I mean.

The only (VERY) bed thing are the file dialogs of gtk apps. They do really look and feel like aliens who speak a totally different language yet unheard by any human beings.

OpenOffice can use KDEs file dialogs! See screenshot:
http://twoday.tuwien.ac.at/pub/files/Bildschirmphoto

Reply Score: 3

RE[3]: All I can say is...
by korpenkraxar on Thu 4th Jan 2007 03:43 UTC in reply to "RE[2]: All I can say is..."
korpenkraxar Member since:
2005-09-10

The only (VERY) bed thing are the file dialogs of gtk apps. They do really look and feel like aliens who speak a totally different language yet unheard by any human beings.

Ha ha ha! Thanx. That made my day :-)

Reply Score: 1

RE: All I can say is...
by cmost on Thu 4th Jan 2007 15:56 UTC in reply to "All I can say is..."
cmost Member since:
2006-07-16

"...Wow!! Looks like I'll be using KDE when 4.0 comes out, something I never thought would happen."

...Wow, and you made that decision based upon a few screenshots of ksysguard and some goofy game? I'm glad to see K developers are spending their energies on the really important elements of the desktop. It never ceases to amaze me how a little eye candy wins out over function or features.

Reply Score: 0

RE[2]: All I can say is...
by JohnFlux on Thu 4th Jan 2007 16:04 UTC in reply to "RE: All I can say is..."
JohnFlux Member since:
2007-01-04

Hey, don't blame the developers.
I've spent a year working on ksysguard improving it on the inside, but that doesn't generate an osnews article.

But a few days work on improving the gui does.

Reply Score: 5

RE[3]: All I can say is...
by cromo on Fri 5th Jan 2007 14:01 UTC in reply to "RE[2]: All I can say is..."
cromo Member since:
2006-06-17

The article was not about ksysguard gui improvements but about SVG support that KDE4 will come. Ksysguard is just an example of SVG usage. I am seriously surprised that you don't get it.

Reply Score: 4

Nice looking resource monitor...
by n0xx on Wed 3rd Jan 2007 18:47 UTC
n0xx
Member since:
2005-07-12

I don't want to start a flamewar or anything but Gnome already uses vector graphics to render certain widgets thanks to Cairo. You guys are playing catch up on this one... :p

http://www.gnome.org/~davyd/gnome-2-14/

Search this page for: Sharp Dressed Man

Having Cairo render the whole interface, now that would be a good thing... the first step towards resolution independence... anyway... nice resource monitor... looking forward to KDE 4

Reply Score: 5

ronaldst Member since:
2005-06-29

@n0xx

I think they have something like Cairo on KDE. Arthur?

IMO KMahjongg is the best example on that page.

Reply Score: 1

elsewhere Member since:
2005-07-13

I think they have something like Cairo on KDE. Arthur?

No, AFAIK Arthur is strictly Qt 4.x; in fact, that may be one advantage that the Gnome/GTK devlopment path has over KDE/Qt, incremental changes and improvements can often be exploited sooner rather than waiting for major version upgrades. KDE 3.x has been hindered at the graphics layer by some of the defficiencies in Qt 3.x that were addressed and improved in Qt 4.x

Having said that, major version upgrades, particularly if they require rewriting API's or libraries, also afford the opportunity to redo things and implement changes in new ways, rather than being constrained by backwards compatibility. But obviously the development cycle is longer and more painful which can penalize a project in the instant-gratification world of OSS rapid release cycles.

Don't know if one way is necessarily better than the other, they both have their advantages and disadvantages. But KDE4 is being designed with an eye to being a stable platform for the next 5+ years, some of the real tangible benefits from the reworked frameworks may not even be evident until 4.1 or beyond. So hopefully the wait will be worth it, particularly with the level of effort the devs are putting in it.

Reply Score: 5

ronaldst Member since:
2005-06-29

No, AFAIK Arthur is strictly Qt 4.x;

Yes. In this context, in regards to the article, we're talking about KDE4 apps.

Reply Score: 4

zsitvaij Member since:
2006-06-14

http://zrusin.blogspot.com/2006/10/benchmarks.html

Having taken a look at these benchmarks, no, using Cairo would be a suboptimal solution, to say the least.

Reply Score: 5

devurandom Member since:
2005-07-06

Interesting. I have not fully understood the reasons for this difference (I know practically *nothing* about GTK, QT and graphic programming, so...) but I wonder:

- how much Cairo can import the improvements of Qt (that is, how much of this performance gain is in the intrinsic architecture of the two projects VS the raw algorithms etc.)?

- could one day the Qt framework be exported to a neutral, external library that both gtk and qt apps can use? or it is structurally impossible to detach it from the graphic libraries?

Naive questions, I know, but...

Reply Score: 2

panzi Member since:
2006-01-22

> - how much Cairo can import the improvements of Qt (that is, how much of this performance gain is in the intrinsic architecture of the two projects VS the raw algorithms etc.)?

It's all in the algorithms. Cairo is already working on a own implementation of this algorithm, which gave Qt a performance of O(n log n).

> - could one day the Qt framework be exported to a neutral, external library that both gtk and qt apps can use? or it is structurally impossible to detach it from the graphic libraries?

I really doubt that this is possible.

Reply Score: 2

superstoned Member since:
2005-07-07

indeed, you can't change a widget (and, in case of Qt, much more) library that easy in an app. Cairo will get faster over time, but it's hard to compete with a toolkit which has a company working on it...

Reply Score: 4

hornett Member since:
2005-09-19

Cairo will get faster over time, but it's hard to compete with a toolkit which has a company working on it.

I appreciate and largely agree what you are saying, but there are various companies contributing to GTK/Cairo at this time.

Nokia immediately springs to mind, as they seem to be doing a lot of work to reduce the amount of floating point calculations done by Cairo (I believe for the Maemo platform).

Edited 2007-01-04 08:50

Reply Score: 1

lliehu Member since:
2007-01-04

>> - could one day the Qt framework be exported to a neutral, external library that both gtk and qt apps can use? or it is structurally impossible to detach it from the graphic libraries?

> I really doubt that this is possible.

Qt 4 has two separate main modules (other modules include qtnetwork, qtxml, ...): qtgui and qtcore. You can link your application to qtcore and choose not to link with qtgui. You may even install just qtcore and not qtgui, but I'm not sure about that. qtcore contains only non-graphic parts of the library.

In Qt 3 this isn't possible.

Reply Score: 2

panzi Member since:
2006-01-22

I just was about to post that link myself. ;)

Reply Score: 1

tristan Member since:
2006-02-01

http://zrusin.blogspot.com/2006/10/benchmarks.html

Having taken a look at these benchmarks, no, using Cairo would be a suboptimal solution, to say the least.


Yes: comparing a narrow test-case that the developer has just been working on against something that (at the time) had received no optimisation in Cairo is clearly a very sensible reason to brand it "suboptimal". Well done!

Cairo now has a re-written tessellator (which is what was being tested in this example). Even the author of this benchmark acknowledges that this will likely show the same logarithmic scaling as his Qt version. Even better, this is available in GTK (and thus Gnome) *today*, rather than having to wait for KDE4 to see the benefits.

And of course, the fact that Cairo is being used in Gecko and OpenOffice means that hackers from those projects will be helping out, fixing performance regressions and corner cases and so on.

Who knows how fast Cairo might be by the time KDE4 comes out?

Reply Score: 4

panzi Member since:
2006-01-22

Good to hear. Even thought I'm using KDE, I also use Firefox, and Firefox could be really faster then it's today.

Reply Score: 1

arielb Member since:
2006-11-15

I must admit I'm a bit concerned and skeptical about firefox using cairo. Once you depend on other people's code, you are stuck to their priorities and agenda.

Reply Score: 2

zsitvaij Member since:
2006-06-14

I must admit I'm a bit concerned and skeptical about firefox using cairo. Once you depend on other people's code, you are stuck to their priorities and agenda.

For what it's worth, I am using the 1.3.10 Cairo snapshot, and Firefox does feel snappier, as do most GTK+ programs.

Anyway, this belongs in the other thread, but these are the latest measurements I found:
http://blogs.gnome.org/view/xan/2006/11/06/0
*EDIT*: two more links I missed, because it seems Xan isn't on the Planet Gnome feed.
http://blogs.gnome.org/xan
http://www.gnome.org/~fherrera/blog//Cairo_faster

Also, these might be interesting:
http://macslow.thepimp.net/?p=101
http://macslow.thepimp.net/?p=100

I do keep updated on developments in the free desktop world, but I'm a wee bit tired of people wanting KDE to adopt some technology or other, based on hype factor.

Edited 2007-01-03 22:24

Reply Score: 5

kaiwai Member since:
2005-07-06

I must admit I'm a bit concerned and skeptical about firefox using cairo. Once you depend on other people's code, you are stuck to their priorities and agenda.

And you know if you write software, you're 'stuck to the operating system developers agendas and priorities' - shock bloody horror!

The difference is, unlike Microsoft, whom if you find a problem they'll tell you to go f*ck yourself (and thus, you're going have to spend the next several months working around those bugs or having to create your own implementation of that said feature) - a developer can download the source, fix the issues as they find them and submit those changes back to the project - they benefit, and all those who rely on Cairo benefit, just as that developer benefits from everyone elses contribution.

Reply Score: 4

arielb Member since:
2006-11-15

I didn't bring up Microsoft or closed source. I support open source OS's (linux is not an OS) where the app developers have a role in its destiny.

Everyone has to deal with the OS they develop for. It's part of the deal. Why complicate things further so that you have less control?

Reply Score: 1

kaiwai Member since:
2005-07-06

Everyone has to deal with the OS they develop for. It's part of the deal. Why complicate things further so that you have less control?

You have as much control over those external dependencies as you want to contribute to those external dependencies projects.

Some are happy with how things are, and simply provide patches for bugs, whilst others want a specific feature added so they work with the developers of that said project to add the feature.

The issue of Microsoft which I bought up had nothing to do with your original post - it was an example of where you as a developer have a dependency on an external supplier and you have little or no control over what is added either by way of bug fixes or features.

I then contrasted that to the open source world where you have the ability to directly add to the project by submitting your own code to fix an issue/bug or add a feature which your project requires/depends upon.

Your application development doesn't occur in a vacuum and as such, you need to maintain open dialogue, not only between you and your dependency's but also those who depend on them and the chain reaction that goes from there.

Reply Score: 2

zsitvaij Member since:
2006-06-14

Cairo now has a re-written tessellator (which is what was being tested in this example). Even the author of this benchmark acknowledges that this will likely show the same logarithmic scaling as his Qt version. Even better, this is available in GTK (and thus Gnome) *today*, rather than having to wait for KDE4 to see the benefits.

The tests are available. Why don't you post the results with the 1.3.10 snapshot?

Reply Score: 3

smitty Member since:
2005-10-13

And of course, the fact that Cairo is being used in Gecko and OpenOffice means that hackers from those projects will be helping out, fixing performance regressions and corner cases and so on.

Actually, one of the stated reasons for Gecko using Cairo was that their developers wouldn't have to help out, thus freeing them up for more web browser related coding. Of course, Gecko provides a lot of test cases to help show where Cairo needs to speed up and maybe some of the developers are working on it now, but I don't expect that to last once Cairo is acceptable for the Firefox project.

Edited 2007-01-03 22:35

Reply Score: 4

arielb Member since:
2006-11-15

No, nobody thinks cairo means mozilla can just hand it over to them. but what it does mean is that mozilla won't be alone in maintaining their own graphics.

I think cairo will be faster. My real problem is that since mozilla didn't come up with cairo, they won't understand bugs that may have to do with cairo but only found in mozilla.

Reply Score: 1

segedunum Member since:
2005-07-06

I don't want to start a flamewar or anything but Gnome already uses vector graphics to render certain widgets thanks to Cairo.

Cairo is an impressive body of work, and is one of those things where people have seen new features and improvements come into Gnome and GTK without a major version number being bumped ;-), but it still needs an awful lot of optimisation and it needs to be made much more programmer friendly.

Reply Score: 3

Plasma project?
by Ford Prefect on Wed 3rd Jan 2007 19:31 UTC
Ford Prefect
Member since:
2006-01-16

I wonder what's the state of the plasma project. The Website seems to stay in an very old, initial state :/

I'm very eager to see some new concepts!

And I'm very interested in wether they also will realize them.


Sadly enough, in the open source world, most times there are no "great concepts", and if they are, they aren't realized; the only new "breaking" concepts I have seen for example in Gnome just were copied from OS X...

Reply Score: 5

RE: Plasma project?
by superstoned on Wed 3rd Jan 2007 22:29 UTC in reply to "Plasma project?"
superstoned Member since:
2005-07-07

well, plasma work is slowly beginning (the new application launcher is an example), but i don't expect to see any great progress soon...

Reply Score: 2

RE: Plasma project?
by ShawnX on Thu 4th Jan 2007 03:29 UTC in reply to "Plasma project?"
ShawnX Member since:
2006-08-04

The website has been updated mostly in the wiki in http://plasma.kde.org/wiki/

But the website used to say the APIs would be ready in 2005, has been dropped since they're pretty busy on with the other bits of KDE 4.0. If you want to help, join #plasma on freenode ;)

Reply Score: 3

RE: Plasma project?
by mcmv200i on Thu 4th Jan 2007 10:30 UTC in reply to "Plasma project?"
mcmv200i Member since:
2006-12-14

Sadly enough, in the open source world, most times there are no "great concepts", and if they are, they aren't realized; the only new "breaking" concepts I have seen for example in Gnome just were copied from OS X...

This is not a valid generalization. Many radical new concepts come from the FLOSS world as well. Just to mention a desktop example: http://en.wikipedia.org/wiki/Project_Looking_Glass

But in some cases these are not the big projects which bring these new things. But you neither can say "Closed-Source stuff is more innovative" nor you can say "FOSS stuff is more innovative". At the moment, both is not true, in general; yet I admit that at the moment apple is by far the most innovative organisation when it comes to usability, blobs and product design.

Reply Score: 1

RE[2]: Plasma project?
by Ford Prefect on Thu 4th Jan 2007 15:35 UTC in reply to "RE: Plasma project?"
Ford Prefect Member since:
2006-01-16

Sure, there are innovative things happening in the FLOSS world! But if they are really done, like Looking Glass (which has a dominant company behind it), they stay to be niche projects.

For example, the Enlightenment Desktop, EFL, etc. is a bunch of very innovative stuff. But it is overlooked by the masses.


Most forces in the FLOSS world are concentrating on competing at the same level with their proprietary counterparts. And it seems most users want exactly that.

For example, if you look at the themes available for Gnome, KDE, etc., the most popular onese are just copying OS X, Windows Vista or even Windows XP (I wonder about that one most!)


I think this is a real problem of the FLOSS world. If it really want's to take off, it has to provide some "killer" apps. It _has_ to be different, in a good way. And I wonder, if this will ever happen.


(Yes I know, there are already many alternative concepts out there, some of them which I also love and use; but these are again niche players)

Reply Score: 2

RE[3]: Plasma project?
by mcmv200i on Thu 4th Jan 2007 16:24 UTC in reply to "RE[2]: Plasma project?"
mcmv200i Member since:
2006-12-14

Maybe - I don't know for sure if this is true - real innovations can't come from commercial projects (if not explicitly marked as "research projetcs" like project looking glass or some other ms projects e.g.) or mass products, because sooner but not later the companies who support the projects want to have a productivly usable outcome.

And I guess this is why KDE and GNOME have to stick to the evolutionary and not revolutionary path; they cannot afford experiments too much. But OS developers only working in their freetime on niche projetcs --- or researchers working on research projects at a big company or at a university --- do have the opportunity to try out something completely new, because they have the right to fail and nobody expects of them to deliver something on time. But the thing which is expected of them is to deliver something radical new.

Reply Score: 1

RE[4]: Plasma project?
by Ford Prefect on Thu 4th Jan 2007 20:05 UTC in reply to "RE[3]: Plasma project?"
Ford Prefect Member since:
2006-01-16

Well, I think real innovations need two things to get popular:

(1) Hard and long work to research them
(2) Aggressive marketing

Both can be done by a company, although it can get very risky! Microsoft, for example, _has_ a very big "research lab", although you wonder, where all the efforts can be seen ;-). Anyway, both are things the FOSS community has trouble to provide.


On the other side, it has the great plus: There is nearly no "risk", you don't get bankrupt if your hobby innovation doesn't work out. But most people wouldn't come over (1) as hobbyist. And even if there are great forces (Linux, Gnome, KDE), these aren't people you can tell "do it that way, I believe it's a groundbreaker"... you have to convince them, and most times, this won't get beyond "common sense", which is most times "let's do it they way others did before, I like that one (because I can't imagine how something else would really look)"...

Reply Score: 2

RE: Plasma project?
by boudewijn on Thu 4th Jan 2007 10:38 UTC in reply to "Plasma project?"
boudewijn Member since:
2006-03-05

Try looking at croquet. Really innovative, open source, but so weird nobody really uses it. Which is the problem with "real" innovations, of course. OS X isn't innovative either, building on concepts and code from the early nineties.

Reply Score: 5

wow
by Techman on Wed 3rd Jan 2007 19:41 UTC
Techman
Member since:
2006-07-21

I LOVE IT ;)

Reply Score: 1

some ideas
by arielb on Wed 3rd Jan 2007 19:43 UTC
arielb
Member since:
2006-11-15

I really liked the mahjong. Some ideas for photorealistic objects:

virtual keyboard
calculator
solitaire game with 'real' cards

be careful though because many real world objects translate to terrible UI.
See
http://homepage.mac.com/bradster/iarchitect/phone.htm

Reply Score: 1

Xaero_Vincent
Member since:
2006-08-18

Beineri writes in his blog about how KDE 4.0 won't have all or even the majority of features planned in the KDE 4 series.

http://www.kdedevelopers.org/node/2600

This is an issue because the KDE 4 series could last five or more years, just like the KDE 3 series. KDE 4.0 has high expectations. If 4.0 turns out only to be minor improvement over 3.5, everyone will very disappointed.

People are expecting a radically new desktop with a largely changed interface in 4.0, not 4.5 in 2011!Something as big as the interface in Windows Vista!

If it takes years before KDE reaches that point, we'll really be in trouble because Microsoft is planning even more interface changes in Windows Vienna.

Reply Score: 3

lord_rob Member since:
2005-08-06

Yeah, except that upgrading from kde 3.5 to kde 4.0 then 4.1 will just be an apt-get (or equivalent) away. You don't need to throw big bucks like you do for Windows updates ;)

Reply Score: 2

ronaldst Member since:
2005-06-29

People are expecting a radically new desktop with a largely changed interface in 4.0, not 4.5 in 2011!Something as big as the interface in Windows Vista!

IMO we've reached the limits of the WIMP desktop metaphor. These days, it's only reimplementing desktop stuff (widgets, finder). Or making apps more integrated into the desktop metaphor like Apple's TimeMachine solution for Backup.

I don't expect much changed to the default KDE desktop except for a few things that look cool.

I'd love to see a desktop like XBox360's UI. It's so cool and usable.

Reply Score: 2

Wintermute Member since:
2005-07-30

How have we reached the limit of WIMP based UI design? None of the main OSs (with the exception fo some minor linux distro) have a comprehensive integrated virtual desktop system. Desktop pager's are pretty decent on linux, but they are only a start, we need organized rule based virtual desktops that are integrated with certain action types. A particular desktop should have settings, layouts, shortcuts suited for certain tasks. And thats just the tip of the iceberg... What about context sensitive file dialogs?

There is so much radical stuff that can be experimented with current desktops.

It's sad to see that companies like Apple and Microsoft have conditioned you to believe that adding a sidebar to your desktop or application such as Timemachine are somehow revolutionary. There is so much more that could be done with current desktops and I am not talking just about the UI itself, think about things like integration with web application. Having a transparent calender that auto syncs with a calendering service and/or auto syncs via blue tooth with your phone would kick ass.

P.S. Apple should release an application called Appleapp, it will be able t launch and you will even be able to close it.

Reply Score: 1

Hands Member since:
2005-06-30

"People are expecting a radically new desktop with a largely changed interface in 4.0, not 4.5 in 2011!"

I don't know who you're talking about here. I don't know anyone who is expecting a radically new desktop with 4.0. I'm expecting an improved desktop with a new framework to build on.

I expect KDE 4.0 to establish a foundation for the future of KDE while being better than what came in the past. KDE 3.5 is better than 3.2 which was better than 3.0. No one can expect developers to reach the full potential of a platform in the first release.

Reply Score: 5

don't like it in ksysguard and run command
by panzi on Wed 3rd Jan 2007 21:04 UTC
panzi
Member since:
2006-01-22

I can't see the point of this for ksysguard and "run command". I actually like it more the old way.

Other then that, it's nice.

Reply Score: 2

Oh my gosh
by Brmbolec on Wed 3rd Jan 2007 21:06 UTC
Brmbolec
Member since:
2005-07-23

I don't know why is that but KDE looks hell ugly. It seems like different kind of people (with different kind of taste) are using KDE/GNOME. I'm MacOSX user, but I have to say I like GNOME more, this KDE thingie works perfect, but on other side it looks weird. Only that Mahjong game looks better on those screenshots...

Reply Score: 0

v No big difference
by vtolkov on Wed 3rd Jan 2007 21:24 UTC
RE: No big difference
by ebasconp on Wed 3rd Jan 2007 22:34 UTC in reply to "No big difference"
ebasconp Member since:
2006-05-09

<Sarcasm>
And the browser... will be implemented on JavaScript or XHTML?
</Sarcasm>

How can you talk about DHTML and internet browsers if you know (I think you know) the browser needs to be implemented using a GUI library [like Qt, KDE extensions, Gtk, wxWidgets or some stuff like this] and needs to implement rendering mechanisms similar to the KDE4 SVG rendering engine?

Criticizing is easy when you are not walking on another else's shoes.

Reply Score: 3

v faffing about
by nzjrs on Wed 3rd Jan 2007 21:41 UTC
RE: faffing about
by borker on Wed 3rd Jan 2007 22:04 UTC in reply to "faffing about"
borker Member since:
2006-04-04

Sigh, lets use a C based, foreign layer rather than the one that comes with the toolkit the entire DE is already built on. Yup that makes sense. And as Arthur is part of Qt, only the subset of trolls that work on both KDE and Qt could be said to be 'fafing about' writing it.

Do you have a link to where the Cairo stuff has caught up? (I'm not saying it hasn't or can't (just like Zach in his blog BTW), just curious if there is any info out there about it).

Reply Score: 5

v RE[2]: faffing about
by nzjrs on Wed 3rd Jan 2007 23:28 UTC in reply to "RE: faffing about"
RE[3]: faffing about
by Chicken Blood on Thu 4th Jan 2007 01:11 UTC in reply to "RE[2]: faffing about"
Chicken Blood Member since:
2005-12-21

Cairo doesn't run on all the platforms that Qt supports. It doesn't exactly serve Trolltech's needs.

And by the looks of things, they're certainly not faffing. One of the graphics layers is well documented and rapidly being improved in response to market forces. The other is Cairo.

Reply Score: 4

RE[4]: faffing about
by nzjrs on Thu 4th Jan 2007 02:12 UTC in reply to "RE[3]: faffing about"
nzjrs Member since:
2006-01-02

One of the graphics layers is well documented and rapidly being improved in response to market forces.

Lol, thats a descritption of cairo. Why? Because firefox is a bigger market force than KDE!

The improvements that the pressure of firefox will bring are guarenteed to be enormous!

Reply Score: 0

RE[5]: faffing about
by Chicken Blood on Thu 4th Jan 2007 02:33 UTC in reply to "RE[4]: faffing about"
Chicken Blood Member since:
2005-12-21

I wasn't referring to KDE, I was referring to all of Qt's paying customers.

Like I said market forces.

Reply Score: 4

RE[3]: faffing about
by borker on Thu 4th Jan 2007 15:46 UTC in reply to "RE[2]: faffing about"
borker Member since:
2006-04-04

'foreign' is not a strawman, KDE depends on Qt. Arthur is part of Qt. Using Arthur does not require an additional dependency, using Cairo would. As with the phonon project, the desire is to avoid dependencies on foreign libraries that have release schedules that are out of step with Qt. As to the 'language war', the desire to program to program in ones 'native' language is hardly unreasonable.

Reply Score: 1

RE: faffing about
by segedunum on Thu 4th Jan 2007 00:33 UTC in reply to "faffing about"
segedunum Member since:
2005-07-06

Sigh. Lets faf about and write Aurthur instead of using Cairo. Sigh

Dear, dear. KDE, and Trolltech who are actually a Linux/Unix ISV, needed something within their own toolkit that works well for their users and customers today.

As for the blog post where zrusin claimed cairo was rubbish

He didn't claim Cairo was rubbish at all. I'm afraid that's just what the benchmarks apparently showed.

check the cairo ML for subsequent discussions and benchmarking - those problems have long been fixed.

No they haven't, and Cairo is by no means developer friendly. I haven't seen any mailing list exchanges or commits that have addressed their performance deficiencies, nor have I seen any benchmarks that prove it.

Edited 2007-01-04 00:33

Reply Score: 5

RE[2]: faffing about
by nzjrs on Thu 4th Jan 2007 02:09 UTC in reply to "RE: faffing about"
nzjrs Member since:
2006-01-02

"apparently showed" because he picked a very narrow and specific test cased designed to show such.

Explain to me how cairo is not developer friendly. It has a very nice API and a plethora of bindings.

The ML is very active and your saying that you have seen no activity or commits addressing performace issues shows that you do not follow it.

See this comment for pointers to some perf increases:
http://www.osnews.com/permalink.php?news_id=16861&comment_id=198344

And you can follow the cairo releases (which include pretty performance graphs) here:
http://cairographics.org/news/

Reply Score: 1

RE[3]: faffing about
by segedunum on Thu 4th Jan 2007 20:04 UTC in reply to "RE[2]: faffing about"
segedunum Member since:
2005-07-06

"apparently showed" because he picked a very narrow and specific test cased designed to show such.

Where is the evidence that he picked a benchmark to show Cairo in that light, and how was the benchmark in any way narrow?

Explain to me how cairo is not developer friendly. It has a very nice API and a plethora of bindings.

In terms of programmer usability, compare this:

http://gnomejournal.org/article/34/writing-a-widget-using-cairo-and...

with this:

http://doc.trolltech.com/4.0/qt4-arthur.html

The ML is very active and your saying that you have seen no activity or commits addressing performace issues shows that you do not follow it.

Comments on mailing lists are not commits, nor are they releases, nor are there any benchmarks proving otherwise - yet.

Reply Score: 5

like brothers fighting
by re_re on Thu 4th Jan 2007 03:25 UTC
re_re
Member since:
2005-07-06

Ahh, all of this is Qt vs Cairo bickering is like two brothers fighting. They piss eachother off and annoy the hell out of eachother but in the end, they are still brothers and when it really counts they are usually on the same side.

I personally prefer qt's framework but it is irrelevant what I think.
The bottome line is that the competition between the two keeps both frameworks striving to become better and more efficient and in the end... it's better for everybody regardless of weather you use kde/qt, or gnome/gtk/cairo.

Reply Score: 2

I love Kde but...
by Drawnstories_studios on Thu 4th Jan 2007 05:12 UTC
Drawnstories_studios
Member since:
2005-12-12

Why! Why must KDE mirror Windows developement?

I mean kde has always looked like a super customizeable explorer.

I like the options, but then now we're back to place where we have to look like vista or die, in other words more vomit coolor schemes and now in special gloss and transparency. yech.

*No Hostility intended*

Reply Score: 2

RE: I love Kde but...
by l3v1 on Thu 4th Jan 2007 14:13 UTC in reply to "I love Kde but..."
l3v1 Member since:
2005-07-06

KDE mirror Windows developement

The only way for me to believe your "No Hostility Intended" line is if you convince me of your ignorance (done). KDE mirroring Windows development ? Why, because some kde skins or menu placements, resemble Windows counterparts ? Ehh, because the resemblance fully stops there, and please notify me the day when the consistent and ubiquitous modularity that the whole kde architecture shows will have anything similar in Windows. After a fairly large number of windows and linux use I found kde to provide a better user experience (of course combined with linux features and tools and applications without which every DE is just a gui without a purpose). Vista almost changed that, almost.

Reply Score: 5

RE: I love Kde but...
by Dark_Knight on Thu 4th Jan 2007 16:55 UTC in reply to "I love Kde but..."
Dark_Knight Member since:
2005-07-10

Both Microsoft and Linux developers mimic one another in one form or another with the intention to attract consumers. If you make a desktop which eases migration from one OS to another it's more attractive to consumers especially those with a LAN that combines Windows and Linux. As for the glossy effect that everyone seems to be using these days KDE will still have the option to change the theme to a more basic look for older lower end systems or people such as yourself that want less eye candy.

Reply Score: 2

thom
by frozen5555 on Thu 4th Jan 2007 07:33 UTC
frozen5555
Member since:
2005-12-27

I guess its partially thanks to thom article on gnome and kde. I see a lot of feedback with information in the news. from both teams.

Browser: Mozilla/4.0 (compatible; MSIE 6.0; Windows CE; IEMobile 6.8)

Reply Score: 1

Redundant ksysguard
by MamiyaOtaru on Thu 4th Jan 2007 08:39 UTC
MamiyaOtaru
Member since:
2005-11-11

What's with the redundancy in ksysguard? Specifically the bar showing current CPU usage. One can simply look at the far right of the graph to see that. What's the point in showing it twice?

See pic if this isn't clear enough:
http://img81.imageshack.us/my.php?image=redundantwz1.png

I'd ask for an option to disable that, but for heaven's sake it should just go. It grates. A better way to do it - if you simply must have the current value highlighted somehow - would perhaps be something like this:
http://img205.imageshack.us/my.php?image=lessredundantut2.png
___
posted at the dot, different audience here though so perrhaps can be repeated

Edited 2007-01-04 08:40

Reply Score: 4

RE: Redundant ksysguard
by frozen5555 on Thu 4th Jan 2007 10:22 UTC in reply to "Redundant ksysguard"
frozen5555 Member since:
2005-12-27

the second is history for the cpu usage.

Reply Score: 1

RE[2]: Redundant ksysguard
by MamiyaOtaru on Thu 4th Jan 2007 13:45 UTC in reply to "RE: Redundant ksysguard"
MamiyaOtaru Member since:
2005-11-11

right.. and the most recent part of that history (the part that's circled) is the current value, which makes the upper circled bit redundant.

Reply Score: 2

RE[2]: Redundant ksysguard
by nutshell42 on Thu 4th Jan 2007 16:54 UTC in reply to "Redundant ksysguard"
nutshell42 Member since:
2006-01-12

A better way to do it - if you simply must have the current value highlighted somehow

your "improvement" is ugly and confusing. Ugly because having part of a history graph bouncing all over the place looks stupid, confusing because that's not how people expect a history graph to behave.

It's hard to discern the current level from the history (it's just a few pixels wide) and ksysguard isn't the only one to have a dedicated graphic for current levels (e.g. the Windows Taskmanager does too).

Reply Score: 3

It looks now more like GNOME
by theuserbl on Thu 4th Jan 2007 21:32 UTC
theuserbl
Member since:
2006-01-10

I have now seen the Screenshots KDE 3.5.5 vs KDE 4 preview.
And I must say, the KDE 4 looks now more like GNOME.

Mostly that now in the Icon-Menu under the icons are also the text - like in GNOME.
That is something I don't like. I try to disable in GNOME-programs the text under the icons, because that take too much space of the desktop. But not all GNOME-programs allowing it to disable it.
Now KDE doing the same thing... :-(

But on the other side, GNOME Mahjongg (especially the new one with the modern SVG graphics) looks much better then the KDE Mahjongg
But the KDE 4 Mahjongg looks now for me as good as the GNOME one.

For me KDE 4 seems to look more like GNOME.

Reply Score: 1

"This article series is a great idea."
by REMF on Fri 5th Jan 2007 10:58 UTC
REMF
Member since:
2006-02-05

Agreed.

< looking forward to more installments on KDE4 progress.

Reply Score: 1

why the backgrounds?
by jziegler on Fri 5th Jan 2007 16:41 UTC
jziegler
Member since:
2005-07-14

I like the idea of writing such progress report articles. I like the idea of describing all displayed elements in SVG.

What I don't get is the background picture in the _graph_. It makes the graph harder to read and I see no other purpose for it, so I thinkg it's useless.

Generally, KDE 4.0 sounds interesting and starts to look good, maybe I'll give it a chance once it's finished and done. Last I tried KDE, some 4-5 months ago, it failed me.

Reply Score: 1