Why Kde should remain as light as possible

Luciano Montanaro mikelima at virgilio.it
Wed Mar 10 10:52:18 CET 2004


On Wednesday 10 March 2004 06:24, Pupeno wrote:
> Just an idea... using PCs as thin clients for, err, big servers is a very good
> way to recycle old PCs... it is good in enviroment like school or ofices, or
> public access or whatever that has a lot of computers altogether.... I didn't
> make the numbers, but I expect that a network set up of thin clients becomes
> cheaper than separated workstation when the number of computers are about 7
> or more. So, there are more than keeping the software light, which is a good
> idea... there are techniches that helps you recycle PCs...

This is a nice idea in general, but you still have to have a powerful server.
I think Marco's idea is to have a distro working well on an aging home computer.

Think a pentium-class PC with 32MB.


> And I even imagine that if the thin clients are not so old, they could become
> part of an Open Mosix cluster with the server making a more powerfull set of
> computers.
> 

Again, this is good if the machine is part of a network. It could be an 
interesting project for a technical school, for example, but does not help
the home user one bit.

On the other hand, I have tried Konqueror 3.2 on my old (1998) laptop 
(64MB, K6-2 350MHz) and it works. Should it have 128MB, it would be perfect.
As it is, after removing unnecessary services, stripping the kernel 
to the minimum and disabling eye candy in KDE, KDE + X manages to almost fit 
in core memory (I have 2-3MB paged out). Using TinyX, and a Qt library optimized 
for size, I think the old laptop would be perfect as a student home computer.
Web browsing, mail and Word processing are within its reach.

About Qt optimization: compiling it with -Os -fomit-frame-pointer -funit-at-a-time
its size drops by quite a bit.

Finally, techniques that help slow and low-memory PCs will have a positive effect
also on more current PCs. See the effect of version scripts on Konqueror loading.
If this technique were used more throughout all of KDE, old PCs as well as new ones 
would benefit. 


> On Wednesday 10 March 2004 00:54, M. Fioretti wrote:
>> Hello,
>>
>> One of the reasons usually mentioned to not care about sw efficiency
>> (here and, sadly, on practically all Linux lists) is the "Moore's Law
>> treadmill": HW would keep getting faster and cheaper so quickly that
>> bothering about efficiency makes much less sense than just buying a
>> new PC every other year.
>>
>> In general, this is a bogus thesis just doesn't make sense in the real
>> world: a world mostly made of people who still have to work months,
>> when they can, to buy any new Pc, and of people who may have the
>> money, but rightly prefer to spend it in a thousand more meaningful
>> ways.
>>
>> There is another reason, and now nobody less than the United Nation
>> University confirms it, recommending to refuse the Moore's law and
>> consumism: on the contrary, the useful lifespan of PCs should be
>> extended as much as possible. Obviously this can happen only if Free
>> SW like KDE is made/kept as modular and light on HW resources as
>> possible (ie quite lighter than today):
>>
>> "Short-Lived PCs Have Hidden Costs"
>>
>> http://www.wired.com/news/technology/0,1282,62562,00.html?tw=wn_tophead_6
>>
>> Ciao,
>>       Marco Fioretti
>>
>> PS: at this point I'm also forced to point out that the non profit,
>> volunteer project in my signature was started two years ago just to
>> provide Free Software that would make following today's UNU advice
>> easier. Any support is welcome: right now, what we need more is ONE
>> TIME help by PHP/MySQL experts to rebuild the website. Please spread
>> the word!


More information about the Kde-optimize mailing list