Forgot your password?
typodupeerror

Comment Correction: funding per 'household' not taxpayer (Score 1) 33

" The BBC is funded via a license fee covered by British taxpayers"

Not exactly correct although some would argue the license fee is a tax.

Every household has to pay the Television License (£145.50 per annum - £12.12 per month) if they have a device capable of receiving terrestrial television signals *and* it is able to receive signals (i.e; antenna connected), or if a person in the household uses the BBC Internet iPlayer service [0].

As well as video services the license funds several national radio services and tens of regional county/city radio stations.

[0] http://www.tvlicensing.co.uk/

Comment Don't blame the victim, blame the perpatrators (Score 1) 383

The point I'm making is that the problem is not Linux, it is the actions of a convicted monopolist - Microsoft - that has so distorted the PC industry that manufacturers are rewarded for only ensuring their products work with Windows.

This starts with hardware components directly implementing Windows protocols, firmware (as in this case) only enabling functionality for Windows, and drivers only released for Windows and no technical documentation released for open-source implementers to work from.

Comment Manufacturer's ACPI DSDT firmware to blame (Score 2) 383

Blame the manufacturer.This is a very common issue that few people realise.

The Advanced Configuration and Power Interface (ACPI) contains the Differentiated Services Description Table (DSDT). The DSDT is byte-code that the active operating system executes within its ACPI pseudo-virtual machine interpreter.

This code provides executable Methods (functions) that perform configuration of hardware components; especially related to power control, enabling/disabling devices (think WiFi radio kill-switch), acting on special platform hot-keys (such as Fn+brightness-{up/down} and much more).

It also contains initial feature configuration methods which enable features based on the operating system identity (OSI). In most DSDTs I've analysed over the last 10 years (and I made a special study of Sony models) the OSI code will only enable the full range of features when the host operating system is Windows.

Generally, when the Windows operating system calls the the OSI methods it passes a text string of the form "Windows XXXX" where XXXX is a year number representing the Windows version. Based on XXXX the OSI method then sets the value of a 'feature' variable, with newer versions of Windows enabling more features.

The default value of the 'feature' variable is usually the lowest possible value. This value is adopted when the host operating system is Linux since the OSI method doesn't recognised the OSI "Linux" string.

To work around this problem Linux kernel provides a command-line configuration parameter that enables it to report itself as another operating system. Using this to 'pretend' to be the latest Windows version recognised by the ACPI DSDT will almost always enable all features optimally.

The kernel parameter is "acpi_osi".

I've used this very recently on a new Asus T300 CHI that failed to enabled power to the USB-connected Synaptics touchscreen digitizer if the the device wasbooted whilst charging from mains.

I also fixed an Ubuntu user's problem last night on a Dell Inspiron where the touchpad would be jumpy and the battery drained very quickly.

In both cases there are 2 steps:

1. Identify the OSI strings the ACPI DSDT supports
2. Add what seems to be the latest Windows version string to the Linux kernel command-line

To find the OSI support strings the easy way (although this may not identify all strings) just extract the strings from the binary table. You could use 'iasl' to decompile the DSDT to source-doe but usually it is not necessary:

$ sudo strings /sys/firmware/acpit/tables/DSDT | grep -C 2 -i windows

The strings of interest will usually begin with "Microsoft" or "Windows". On the Asus T300 CHI they are:

Windows 2009
Windows 2012
Windows 2013

The latest is "Windows 2013".

On an older Dell XPS M5130 they are:

Microsoft Windows
Microsoft Windows NT
Microsoft WindowsME: Millennium Edition
Windows 2001
Windows 2006

The latest for the M1530 is "Windows 2006"

Edit the boot-loader's configuration to add the kernel command-line option to any existing settings. For GRUB2 that'll be editing /etc/default/grub:

GRUB_CMDLINE_LINUX="acpi_osi=! acpi_osi=\"Windows 2013\" "

Notice that we need to escape (\") the embedded double-quotation marks inside the variable definition in order for the quote marks to be written to the kernel command-line itself. After saving the file do:

$ sudo update-grub

which writes the changes to the GRUB2 boot-time configuration script: /boot/grub/grub.cfg

Each linux kernel boot entry (lines beginning "linux ...") will include the parameter similar to this:

linux vmlinuz .... acpi_osi=! acpi_osi="Windows 2013"

acpi_osi=! first deletes all recognised OSI strings.
acpi_osi="Windows 2013" sets the only recognised OSI

You can also test the option before adding it permanently by editing the boot-time boot menu entry. For GRUB2 that involves holding down the Shift key during the firmware POST until GRUB2 is loaded, which will cause GRUB2 to display its boot menu. From there highlight the entry to wish to modify, press 'E' to edit it, navigate down to the line beginning "inux ..." and add to the end of it:

acpi_osi=! acpi_osi="Windows 2013"

Obviously replace my OSI string with the one you've identified (or guessed) then press Ctrl+X or F10 (as noted in the on-screen hints) to immediately boot with this modified configuration.

You can check the current boot's active parameters with:

$ cat /proc/cmdline
BOOT_IMAGE=/vmlinuz-4.4.0-21-lowlatency root=/dev/mapper/VG02-rootfs ro no_console_suspend initcall_debug debug acpi_osi=! "acpi_osi=Windows 2013" splash vt.handoff=7

If you have 'debug' level kernel logging enabled you might also see some ACPI OSI related messages in the dmesg kernel log:

Apr 24 04:30:07 T300CHI kernel: [    0.106325] ACPI: Added _OSI(Windows 2013)

Comment Any other symptoms? (Score 5, Interesting) 207

That sounds like the reading equivalent of 'jumping to conclusions' in spoken conversation where the subject believes they already know how a sentence is going to be completed and jumps in with an answer before the speaker has finished.

How did you combat it? Does word-counting help? Does it affect both printed and electronically displayed text? Do you get any other symptoms like headaches?

It reminds me of some of the symptoms of Visual Stress a.k.a. Meares-Irlin syndrome [0].

I helped a friend many years ago (2002) who was thought to be dumb because he seemed unable to absorb written material and after 1/2 a page would switch to light skim-reading ("speed reading") and/or distract himself in any way possible. Being questioned on the material later he would be unable to answer many questions due to skimming over the material, leading to the 'dumb' tag.

He would also sometimes complain of severe headaches that could last days. Since childhood parents, teachers and doctors had tried to find a cause and subjected him to all sorts of tests with no result.

One day whilst we were focused on some programming he complained of a headache. Being the first time I'd witnessed his symptoms I asked him to describe exactly what he was experiencing. It turned out the printing would begin to swim around and blur in and out of focus and get worse the longer he tried to focus on it. He'd never been asked this question before and had assumed everyone experienced this and had not mentioned it.

After some research I discovered Professor Arnold Wilkins at Essex University, U.K., had developed a diagnostic test that identified the cause and possible counter-measures.  Meares-Irlen syndrome is a visual acuity abnormality that can be partially or fully re-mediated with the use of colour filters, with each sufferer needing filters tailored to them - rather like a lens prescription for glasses.

We visited the university and my friend undertook the test and immediately noticed an improvement once the correct colour filter was identified. These tests were done whilst placing permutations of coloured transparencies over printed material (black text on white paper).

As a result I wrote a program that detected and applied the correct colour overlay to the computer screen and it worked as well as the transparencies but the colour required was quite different - due to the differences between reflective and transmissive light.

[0] "Colour in the treatment of visual stress" http://www.essex.ac.uk/psychology/overlays/
[1] "READING THROUGH COLOUR" http://www.essex.ac.uk/psychology/overlays/book2.pdf

Comment Ubuntu != ubuntu-desktop (Score 1) 23

The problem for Ubuntu users and others is the confusing use of the "Ubuntu" project name to also denote the default GUI desktop flavour, the "ubuntu-desktop" virtual package-name, or the installer ISO name.

This article is talking about Ubuntu as guest in virtual machines where the install images are generally from the ubuntu-cloud [1], Canonical partner-images [2] or ubuntu-server ISOs [3].

Ubuntu GUI users are probably only familiar with the GUI flavour names, as in:

$ apt-cache search -n '.*buntu-desktop'
ubuntu-desktop - The Ubuntu desktop system
mythbuntu-desktop - The Mythbuntu standalone system
edubuntu-desktop - educational desktop for Ubuntu
kubuntu-desktop - Kubuntu Plasma Desktop/Netbook system
lubuntu-desktop - Lubuntu Desktop environment
...
xubuntu-desktop - Xubuntu desktop system

[1] https://cloud-images.ubuntu.com/
[2] https://partner-images.canonical.com/core/xenial/
[3] http://cdimages.ubuntu.com/ubuntu-server/

Slashdot Top Deals

We can predict everything, except the future.

Working...