19

This idea cannot be distribution specific, since it is not simply a question of not installing a few packages.

*I have personally disabled several services/packages and run KDE 4.4.2 on an old machine and got down startup memory usage to 200mb.

*I have disabled Strigi, Nepomuk and Akonadi (kwriteconfig --file kres-migratorrc --group Migration --key Enabled --type bool false)

*I uninstalled Kmail, Korganizer, Kontact, KDE PIM (I live off gmail + skype)

*Oxygen/Air theme are too processor intensive - used some other themes to reduce memory + processor usage.

It would be nice if a slim version of KDE is available without semantic desktop dependencies. Today, I see that if I try to uninstall nepomuk or akonadi packages, there is a dependence of plasma on it.

This could mean build packages (qmake scripts, etc.) that leave out all these dependencies. Maybe that could mean some other libraries can also be left out and it will make the desktop run much faster.

I hate the fact that XFCE/Openbox sells itself as lighter than KDE - KDE can be made much, much lighter by removing some non-critical components, yet retaining the end user experience which is uniquely KDE.

In today's time of netbooks (which especially are exploding in Asia), I dont see KDE _at all_. All I see are Netbook versions of Gnome, Openbox and XFCE.

Its time to take it back!

I agree with this wholeheartedly. We have variations of KDE being developed for phones, netbooks, this that and the other; Why not develop a KDE variation specifically for low-end machines, but not necessarily netbooks?

If we do this, we can strip out so much, and still allow extras via kgethotnewstuff. For example, keep the plasmoids down to a select few, remove many of the extra programs, and reconfigure existing systems.

I don't know if it would fly, but we could strip out 90% of the compositing features (leaving only the option for general acceleration), nepomuk, strigi, duplicate programs with plasmoid relatives (calculator, virtual keyboard), include only one default plasma theme (a very light variation on air, optimized for non-compositing desktops?), etc.

I also (will) plus this because the netbook remix is built specifically for small screens with a different interface, and it's still not actually all the slimmed-down. Though I think if this was done, it would be a good idea to start by basing KDE (LWSC?) on the netbook edition.

Some might say it's a distros decision to do this or not.
But I still think it's should be provided by KDE.
Especially there are so many other KDE (plasma etc.) versions.
I'm voting it up, but on the other hand I believe some probably will down vote it.

I'm not voting it down, but I do have some doubts. While it sounds a good idea, it would necessarily have to ship without nepomuk, I think, and without nepomuk, the future of quite a few apps would be broken. To ship a version that doesn't in any way depend on or use nepomuk would need a huge investment of developer time, and I doubt if there are developers to spare to do this. OTOH, if developers able and willing *could* be found, it would be popular with owners of older hardware, as well as netbooks.

This is definitely distribution-specific. There is no reason you have to install the KDE pim applications, for instance (akonadi included). You have to install nepomuk, KDE cannot work without it, but there is no reason you need to have strigi enabled, nor any reason you need to use (or even install) the oxygen them. I don't see any benefit from KDE releasing a whole other version when all it takes is uninstalling a couple of packages and changing a couple of settings.

I think there is discussion in kde-core-devel mail-list about installation profile specially for limited device.

See here:
http://lists.kde.org/?t=127107095800007&r=1&w=2

You have to install nepomuk, KDE cannot work without it

What do you mean? I've disabled nepomuk service autorun and it still works...

If someone chooses they can easily make a lite kde distro. For example in Arch there are some minimal kde packages that leave you with a very bare kde desktop (not even konsole or dolphin installed!). Sidux also has a lite-kde version.

I personally think there should be a light kde distro that is easy to install so that people who moan about kde bloat can see what kde has to offer.

@annew: While it is important for future apps to use nepomuk, I think that those apps should still be able to function without it. For example if a windows user wishes to install amarok they shouldn't have to install nepomuk if they don't want to. I think having optional and minimal dependencies is crucial for kde to break into other OS markets.

@Tenno
I think you do not have to activate indexing, but applications like KMail or even Kopete or Plasma need nepomuk- and strigi-libraries.

Well, I haven't looked the source code so I don't know, but that's possible.

I think this is a good idea, but we should keep kdepim if possible.

IAM very interested in getting the current KDE 4.4.2 to run at 200 Meg in memory. Can you please post more information on how you were able to do this.

I am currently working on a Netbook based distribution and we focus somewhat on devices with smaller hard drive foot prints. Like the EeePC 2G Surfs. We run in a live environment compressing our 2G worth of packages too 700M and then decompressing files etc.. on the fly to a ram disk and saving any changes to the rest of the 1.3G drive. With KDE running at around 400M in memory and the netbook only having 512 we run into major issues opening files or anything that gets placed into memory that's any more then 100M in fact it crashes the KDE environment and puts the use back at kdm. We did not have this issue with the KDE 3.5 series and we don't have it with XFCE. So any input more then what you've already gave would be very much appreciated. Of Course I will vote for this because I feel memory usage has become bloated there needs to be a way or an option to turn off the bloat that causes it.

Various arguments about KDE 4 not being as memory intensive as 3 or just turn of stigi have done us no good. This might look great on the forums for trying to explain away a users complaints but memory usage was something we relied on and we have seen increase considerably with 4 so much so it affects our decision to use KDE 4 in the future unless some slim option is created.