World: r3wp

You get a sort of family box, such as a D-Link, that handles the 
connection method and provides a standard ethernet hub for the whole 
house

Routers you can buy separately all have PPPoE capability for a number 
of years

I was choosing the most cheap providers (but not the most cheap Internet 
connection badnwidth).

So when that changes in Russia, Syllable will boom there ;-)

In the Russia, there is normally one computer per family, and in 
some cases a few computers per family.

Heh. qemu emulation of Syllable under gNewSense is 50 times faster 
than under latest LUBUNTU. Probably because kvm module fails to be 
active under  LUBUNTU.

This speed is visible by mouse cursor responsivity.

*responsiveness

Another guess is that LUBUNTU was 64-bit OS emulating 32-bit Syllable. 
The gNewSense is 32-bit.

Yes, when I tested several years ago, the kernel module made emulation 
roughly twenty times faster

Does Syllable recognise your QEmu video emulation? It should on newer 
QEmu versions, probably as VMware graphics

I don't have VMware, and I don't usually use pirated software (i.e. 
no way to run VMWare).

Now Syllable is in VESA mode at QEMU. QEMU is set to one of the two 
QEMU's possible graphics cards: "Cirrus Logic GD 5446 PCI VGA video 
card". The other (unselected) video card possible is "stdvga (stdbga/vesa)".

vmware server is free is it not?

free to use ..

I don't know.

http://downloads.vmware.com/d/info/desktop_end_user_computing/vmware_player/4_0#open_source

Unsupported

I consider the following features of Syllable as its strongest points: 
1) Simplicity, and 2) having a precise target: desktop OS. Targetlessness 
makes a bad thing to operating systems... (Though I didn't think 
too much on strongest points, and I don't know much.)

Internal simplicity, I mean.

Kaj: could you draw a total Syllable high-level architecture overview 
diagram (i.e. levels and subsystems)? Is such a diagram present at 
Syllable docs?

Such a diagram could be published on a Syllable development site, 
in the About section.

As in: bird view diagram.

Kaj: What is better --- to take latest sources of Syllable, or to 
take stable network sources? I am hesitating

That's re: network stack

I think I will now work with stable ones

If you have only one Syllable system, it's best to use the stable 
0.6.6 release

However, if you can have multiple installations, development work 
is best done on the development build. The kernel headers have changed 
a lot, which affects the source code of system modules

The development build lacks some parts, though. Most notably cURL, 
so that the Webster browser doesn't work

If you want to install a development build, you may want to wait 
a week. I intend to have a new one for the Syllable Conference coming 
weekend

We don't have a Syllable architecture diagram. It would be nice to 
have one, but unfortunately, I have many other priorities. However, 
much BeOS documentation applies to Syllable, especially the overall 
architecture. Haiku keeps a lot of such documentation, such as the 
BeBook

I prefer to work with stable version first, to not encounter clashes 
with newborn bugs. If I get to anything working, I would like to 
merge it with the development version. So I will not currently install 
a development version. I have one computer with four hard drives 
and multiple partitions, one QEMU Syllable installation, and one 
native HDD Syllable installation.

newborn bugs --- i.e. those bugs caused by other developers.

I've tested Haiku. 1) It does not have pppoe out of the box, and 
afair on its forums there is an unfulfilled request for pppoe. 2) 
Haiku uses VESA mode on my machine, supports more modes than Syllable, 
and is MUCH MUCH faster re: windows dragging with mouse than Syllable. 
Haiku reports: VESA version = 3.0, capabilities 1. Haiku allows for 
1280x1024x32 mode while Syllable only allows 1280x800 mode and shows 
a black screen in 1280x1024 mode.

Haiku is fast in VESA mode, Syllable is extemely slow re: windows 
dragging with mouse in vesa mode.

evgeniy, have you looked to minix3 too?

not yet. at my plans there's plan9 and inferno, too

looked at inferno OS. All is bytecode-compiled, therefore rubbish.

Plan9 is interesting.

I think Plan9 is now forked into Front9.

Minix3. wikipedia: "The main uses of the operating system are envisaged 
to be embedded systems and education, such as universities or the 
OLPC XO-1 laptop.[3]" --- not very intreresting goals for me.

plan9front is the name, sorry

My preference are fast FOSS desktop OSes.

RTOS are also interesing for the abovementioned intent but currently 
are way above my head.

microkernel architecture and selfhealing. perhaps the real hurd? 
;-)

self-healing is one important property to study; yes.

but microkernels are unavoidable. everything will have them with 
time.

Processors will evolve to provide a fast zero-copy read-only safe 
access to kernel's memory for usermode processes.

Yes, Haiku's VESA driver is known to be fast. It could be interesting 
to port it to Syllable

BTW What areas are strong in Syllable over Haiku and in Haike over 
Syllable?

They're quite alike and their development state is also quite alike. 
I can't really list all the details, as I only test Haiku once every 
few years or so, when they make a new alpha release